Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Distinct psychological profiles in multiple sclerosis : relations to site of lesion and neuropsychological… Boyle, Eleanor A. 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1992_spring_boyle_eleanor.pdf [ 2.01MB ]
JSON: 831-1.0086673.json
JSON-LD: 831-1.0086673-ld.json
RDF/XML (Pretty): 831-1.0086673-rdf.xml
RDF/JSON: 831-1.0086673-rdf.json
Turtle: 831-1.0086673-turtle.txt
N-Triples: 831-1.0086673-rdf-ntriples.txt
Original Record: 831-1.0086673-source.json
Full Text

Full Text

DISTINCT PSYCHOLOGICAL PROFILES IN MULTIPLE SCLEROSIS:RELATIONS TO SITE OF LESIONAND NEUROPSYCHOLOGICAL FUNCTIONbyELEANOR A. BOYLEB.A. University of Chicago, 1987A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDepartment of NeurosciencesWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIA1992Eleanor A. Boyle, 1992In presenting this thesis in partial fulfilment of the requirements tor an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)____________________:‘Department of S C\The University of British ColumbiaVancouver, CnadaDate /\\DE-6 (2/88)ABSTRACTAn important question in multiple sclerosis (MS) iswhether personality changes result from theneuropathological process or, alternatively, are normalpsychological responses to the stress of living with thedisease. The purpose of this study was to determine (1) ifempirical support could be found for personality profilesdescribed in the literature and (2) if these profiles wererelated to neuropathology. The subjects for this study were99 MS patients with mild physical disease, and 56 well-matched normal controls. In addition, a validation studywas done on positive results, using 43 MS patients and 20controls.A review of the literature suggested four commonprofiles in MS patients, namely denial, exaggeration ofsymptoms, ‘Depression’ (distinct from psychiatrically—defined major depression), and distress concordant withphysical disability. These profiles were operationallydefined by three variables measuring (i) objective clinicaldisability (Kurtzke Expanded Disability Status Scale); (ii)the patient’s perception of symptoms; and (iii) thepatient’s level of psychological distress. These threevariables were analyzed using Ward’s method of clusteranalysis, which yielded four groups consistent with thehypothesized profiles. Subsequently, identical results wereobtained on the validation study.iiTo determine if profile membership was related toneuropathological processes, membership in a psychologicalprofile was correlated with lesion site and cognitivefunction. Site—by—site lesion analyses revealed that‘Depressed’ patients had more pathology in one rightparietal lobe site than did members of other groups. In thevalidation study there was a non—significant trend(p < 0.09) supporting this relation. The remaining sites(26/27) showed no differences, therefore, with the exceptionof the ‘Depressed’ group, no support was found for apathological basis for psychological profiles. Hence, onewould infer that in the mild stages of MS, profiles arereactive responses. The profile groups were also comparedon number of lesion sites and did not differ, suggestingthat profile does not reflect stage of biological disease.Analyses of neuropsychological test results indicatedthat only the ‘Depressed’ group with the right parietallesion had cognitive impairment in comparison to the othergroups. This finding was not replicated on validation.The current study provided empirical support fordistinct psychological responses in MS, but for denial andsomatic exaggeration no evidence was found for apathological basis. These data will be useful toprofessionals working with MS patients, and may havetherapeutic implications.iiiTABLE OF CONTENTSPageABSTRACT iiTABLE OF CONTENTS ivLIST OF TABLES ViLIST OF FIGURES viiLIST OF APPENDICES viiiACKNOWLEDGEMENTS ixCHAPTER ONE:INTRODUCTION AND OVERVIEW 1Purpose of the Study 1Difficulties Inherent in the Problem 2Psychological Change Has Long Been Reported 4Psychological Profiles in MS 10Hypotheses Underlying the Study 13General Methodology 15Potential Contributions of the Study 18CHAPTER TWO:BACKGROUND TO THE STUDY 20Organisation of the Chapter 20Psychological Aspects of MS:Several Distinct Responses 21Response Types Identified in the Literature 30Psychological Problems:Cause or Result of the 35Psychological Problems as a Result of the Disease:Reactive or Organic’ 37The Study: Conceptually 40The Study: Rationale for Methodology 47CHAPTER THREE:METHODS 53Subject Selection and Characteristics 53Test/Assessment Procedures 54Development of Scales 56Cluster Analysis 58Relations Between Groups and Measures of Disease . 61ivCHAPTER FOUR:RESULTS 72Subject Characteristics 72Derived Measures: Sx and Ds 75Cluster Analysis 76MRI Analysis 81Neuropsychological Analysis 86Post Hoc Analyses 93CHAPTER FIVE:VALIDATION STUDY: METHODS 101Subjects 101Sx and Ds Scales: Reliability and Validity 101Cluster Analysis 102MRI Analysis 103Neuropsychological Analysis 103CHAPTER SIX:VALIDATION STUDY: RESULTS 105Subjects 105Sx and Ds Scales 108Cluster Analysis 108MRI Analysis 110Neuropsychological Analysis 112CHAPTER SEVEN:DISCUSSION AND CONCLUSIONS 116Methodological Considerations 116Findings 128Implications 131Further Comments 135Contributions of the Study;Suggestions for Future Research 142APPENDICES:APPENDIX A 144APPENDIX B 148APPENDIX C 151REFERENCES 156vLIST OF TABLESPageTABLE 1 Neuropsychological tests used: dimensionsof cognition measured by each test 51TABLE 2 Neuropsychological tests used: gray matterlesions thought to be associated 52TABLE 3 Characteristics of MS subjects 73TABLE 4 Comparison of MS subjects with normalcontrols on demographic variables 74TABLE 5 Sx- and Ds-scale analyses 77TABLE 6 Profiles do not differ significantly onnumber of sites with lesions 82TABLE 7 Percentage of subjects with lesions in eachbrain area 84TABLE 8 Gray matter-white matter (GM-WM) junction ofthe right parietal lobe: chi-squared analysis . 85TABLE 9 Neuropsychological tests on which ANOVA5showed significant differences among groups ... 88TABLE 10 Correlations between tests on which ANOVA5showed significant differences among groups ... 94TABLE 11 Comparison of groups on total number ofsites with lesions, for groups subdividedin reference to right parietal lesion 97TABLE 12 Comparison of two ‘Depressed’ groups,other MS subjects, and controls, onBeck Depression Inventory (BDI) 98TABLE 13 Validation study:characteristics of MS subjects 106TABLE 14 Validation study: comparison of MS subjectsand normal controls on demographic variables . 107TABLE 15 Validation study: Sx- and Ds-scale analyses .. 109TABLE 16 Validation study: Profiles do not differsignificantly on number of sites with lesions 113TABLE 17 Validation study: GM-WM junction of the rightparietal lobe: chi-squared analysis 114viLIST OF FIGURESPageFIGURE I Predicted psychological profiles 16FIGURE II MRI lesion-location analysis:hypothesized ideal data 64FIGURE III Potential profile groups forneuropsychological analysis 69FIGURE IV Neuropsychological data analysis:hypothesized ideal data 70FIGURE V Results of cluster analysis:psychological profiles in MS 79FIGURE VI Results of Tukey’s tests forneuropsychological tests on which ANOVAsshowed significant difference among groups89FIGURE VII Graphical comparison of groups onneuropsychological tests on which ANOVAsshowed significant differences 92FIGURE VIII Validation study:Results of cluster analysis 111FIGURE IX Schematic illustration of overall results127viiLIST OF APPENDICESPageAPPENDIX A Correlational studies which havebeen done in MS, between psychologicalvariables and physical/disease variables144APPENDIX B MRI lesion-location analysis:hypothetical distributions of subjectswithin groups, and associated p—valuesAppendix B.l 148Appendix B.2 149Appendix B.3 150APPENDIX C MMPI items comprising Sx and Ds scales . 151viiiACKNOWLEDGEMENTSMany thanks to the members of the dissertationcommittee. Dr. Donald Paty, Dr. Joel Oger, Dr. Robert Conry,Dr. David Li, and Dr. Athanasios Zis have given ongoingencouragement, and thoughtful comments on the project and onthis written work. Special thanks to my advisor, Dr.Campbell Clark, for patience and good humour, for rigor inmethodology, and for determined iconoclasm.I would like to acknowledge the many individuals whohave helped me on the journey through graduate school,including Dr. Pat McGeer, Dr. John O’Kusky, DebbieWeinstein, Joane Sunahara and Kim Good. To all theneurological patients who volunteer for our studies, thankyou for your co—operation and optimism. I am grateful toSally Adams for her work with subjects, and to Dana Sair forher thesis—writing expertise.To the Medical Research Council of Canada, I am verygrateful for Studentship funding.ixCHAPTER ONEINTRODUCTION AND OVERVIEWPurpose of the StudyThere is longstanding evidence that multiple sclerosis(MS) patients may exhibit psychological abnormalities, bothcognitive and affective. However, the relation betweenpatients’ psychological status and the disease process isnot clear. Simply, it is not known to what degreebehavioural abnormalities result from the organic diseaseprocess or, alternatively, are secondary reactive responsesin individuals faced with a chronic, debilitating disease.It has been difficult to differentiate between, on the onehand, cognitive and affective changes arising directly fromphysical disruption to the nervous system and, on the otherhand, individual coping or accommodation responses to suchdisruption. Whereas caution must be exercised in theapplication of dichotomies such as “organic” vs. “reactive”,because there tends to be overlap between categories ofhuman making, the distinction is nevertheless an importantone in medicine in general, including in MS. The purpose ofthis study, then, was to attempt to differentiate betweenorganic and reactive changes, by assessing (1) whetherpatients show distinct psychological profiles, (2) whetherpatients with any given profile share common lesion sites1suggesting an organic basis to the profile, and (3) whetherpatients in a given profile have similar cognitive problemssuggesting common lesion patterns.Difficulties Inherent in the Problem: Past and PresentThe difficulties in distinguishing between organic andreactive abnormalities in patients with MS exist for severalreasons. Firstly, to know whether behaviours areorganically—determined, we must have precise information onthe anatomical location of lesions and the degree ofpathology within an individual brain. Such information hasbeen difficult to obtain and analyse. The distribution ofMS lesions is highly variable from patient to patient andthe ability to image white matter lesions has been limiteduntil recently. Computerized axial tomographic (CT) imaginghas been used, but is limited by poor contrastcharacteristics (Bydder et al. 1982) and by the potentialinvasiveness of contrast procedures. High—volume double—dose delayed CT (HVD CT) is more sensitive than other CTmethods, and is particularly useful for showing blood brainbarrier disruption. However, the recent introduction ofmagnetic resonance imaging (MRI), which does not necessarilyemploy contrast material or radiation, has provided moreaccurate identification of white matter lesions with noknown risk to the subjects. MRI has a sensitivity of2perhaps 10 times that of CT in detecting MS lesions (Patyand Li 1988)Secondly, our understanding of functional neuroanatomyis based on the contribution of gray matter regions tobehaviour, and at present there is little theoreticalfoundation for understanding the effects of white matterpathology on behaviour. Therefore, it has been difficult todraw inferences concerning brain/behaviour relationshipsbased on white matter lesions. Because MS has beenconsidered primarily a somatic disorder, a third difficultyin distinguishing between organic and reactive psychologicalproblems is that patients’ mental states have not beensystematically documented in the past. Fourthly, MS is arelapsing-remitting disease with a notoriously variablecourse. Hence, numerous variables may change inunpredictable ways from exacerbation to remission. Finally,there are the ever—present difficulties of correlating anybehaviour with a neuroanatomical structure or region,considering the complexity of central nervous systemconnections. Brain lesions can produce numerous kinds ofbehavioural abnormality, particularly if the lesions are inhigher cortical centers. Such abnormalities can include notonly losses, but sudden appearances of simpler behaviours(Jackson 1884, reprinted in Taylor 1958). Kolb and Whishaw(1980) have suggested that brain lesions can have at leastthree different effects on behaviour: loss of function,release of function, or disorganisation of function.3Because of the brain’s complex interconnections, a lesioncan have multiple effects — producing loss in one behaviour,for example, and disorganisation in another behaviour.Psychological Change Has Long Been Reportedand its Origins DebatedSince before the turn of the century, investigatorshave reported psychological problems in MS and debatedwhether the phenomena were primary or secondary to thedisease. Charcot first reported psychological abnormalitiesin patients with what he then termed sclerose en plaques.According to Charcot, patients often displayed intellectualand emotional deficiencies, showed indifference to theirphysical problems, and frequently laughed or cried withoutvisible cause (Charcot 1877).During the 20th Century, numerous authors havepostulated the existence of an MS personality, namely aconstellation of premorbid psychological traits or problemspredisposing some individuals to the disorder. Severalstudies described patients as immature, anxiety—ridden, andotherwise emotionally abnormal (Grinker et al. 1950;Langworthy 1950). Inman (1948) submitted that MS “is asomatic reaction to intolerable mental conflict” (p. 154).Philippopoulos et al. (1958) reported that MS patients hadexperienced unhappy childhoods and that they displayedemotional problems which may have contributed to the4development of the disease. In the assessment of Grinker etal. (1950), as MS progresses the patient “may neurologicallyactually become the infant that he has always beenpsychologically” (p. 459). However, implications thatregression necessarily reflects psychological immaturityfail to recognise that both cognitive and physicalregression have been noted in other neurological diseases aswell as in aging (Lezak 1983). Hughlings Jackson used theterm “dissolution” to refer to the result of higher corticaldamage in which individuals lose their most sophisticatedbehaviours and regress to simpler, more—primitive ones(Jackson 1884, reprinted in Taylor 1958).Recent reviewers (VanderPlate 1984; Peyser and Poser1986) have commented unflatteringly on some of the earlystudies in MS psychology, which they said were too oftenbased on biased samples, lack of objective data, and lack ofappropriate controls, and which drew conclusions unsupportedby the data. In the view of Peyser and Poser (1986), suchstudies unfairly perpetuated negative descriptions of MSpatients. “The existence of a contributory relationship,possibly through the immune system, is an entirely plausiblenotion, but such a formulation most certainly does notdepend on a premorbid hysterical personality style” (p.383)Personality studies using the Minnesota MultiphasicPersonality Inventory (MNPI) have tended to reinforce thenotion that MS patients exhibit neuroses and even5psychopathology. Numerous authors have reported that MSpatients score high on the MMPI’s Hypochondriasis (Hs),Depression (D) and Hysteria (Hy) scales, the “neurotictriad” (Canter 1951a; Baldwin 1952; Gilberstadt and Farkas1961). Elevated scores have also been reported on theSchizophrenia (Sc) scale, which is sensitive to bizarrefeelings and peculiar body dysfunctions among otherproblems. Researchers have not always interpreted theelevated scores as indicative of an MS personality. Canter(1951a) suggested instead that the high scores represented acoping style. Nevertheless the data have implicitly seemedto support an “MS personality” hypothesis. One problem isthat instruments such as the MMPI were designed for, andvalidated on, normal and psychiatric population samples.Therefore individuals who genuinely suffer from physicalillness may have spuriously elevated scores on scales suchas Hypochondriasis, merely because they are endorsing itemsrelating to their physical difficulties. This confound hasbeen pointed out not only for MS but for rheumatoidarthritis (Pincus et al. 1986) in which patients also scorehigh on Hs, D and Hy scales of the MMPI. Prigatano (1987)comments that the MMPI can be misleading if used onindividuals with organic brain damage. MS patients’elevation on neurotic scales (Hs, D and Hy) as well as theSc scale may result primarily from endorsement of itemswhich reflect true physical symptoms and related healthconcerns. Deletion of MMPI items directly related to MS6syxnptomatology has been shown to produce profiles muchcloser to normal (Marsh et al. 1982). These findingsunderscore the hazards of making inferences aboutpersonality in MS patients from standardized tests.Specifically, some of the characteristic signs and symptomsof MS are also signs and symptoms of diagnosable psychiatricdisorders. For example, assessments of depression are oftenbased on somatic complaints such as fatigue and physicalweakness, two common physical symptoms in MS. Similarly,many signs and symptoms of MS are of the type associatedwith hysterical conversion reactions. Accordingly, early—stage MS is often initially mistaken for psychiatric illness(Skegg et al. 1988).Counter to the views of the “MS—personality” theoristsis a school of thought which asserts that MS psychologicalproblems are secondary and therefore a response, rather thana contributing factor, to the disease. This approach hasprompted questions regarding the origins of such problems,which could broadly be typed as either biological/organic,or situational/reactive. It has been suggested thataffective problems can be reactive, such as in MS patientswho have spinal cord lesions only (according to clinicalassessment) whose depression nevertheless increases withdisability (Mclvor et al. 1984). However, as noted byBerrios and Quemada (1990), a distinction between cerebraland spinal cases is clinical and not neuropathological; evenin spinal cases there may exist brain lesions which are7neurologically but not psychiatrically silent.Nevertheless, it has also been shown that depressionexhibited in MS is akin in incidence and severity to thatobserved in muscular dystrophy, which is also a progressivedisease leading to paralysis but not involving the brain(Surridge 1969). Correlational studies have shown thatdepression is more severe during exacerbations than duringremissions (Cleeland et al. 1970) and that depression isrelated to degree of disability (Baretz and Stephenson1981); however, such findings could be evidence either foran organic or a reactive cause. For example, if depressionincreases as the disease produces more and more disability,the increased depression could have resulted either fromadvancing organic pathology or, alternatively, fromaccumulated stress. In summary, current reviews of theliterature do not vindicate mid-twentieth-centurycontentions for a typical premorbid MS personality, butsuggest rather that psychological abnormalities can ariseeither directly from the pathological process, or from aninterplay of the pathology and the personality of theindividual involved.Though much research on psychological aspects of MS hasinvolved affective changes, cognitive dysfunction in MS isincreasingly being documented. For decades, it was believedthat such dysfunction might be a mild and relativelyinconsequential manifestation of the disease, or might beseen only in its late stages. It was also unclear whether8apparent cognitive difficulties merely reflected patients’motor and sensory problems. However, there is mountingevidence that cognitive changes are a persistentmanifestation in MS (Rao et al. 1984; Rao et al. l989d;Beatty et al. 1988, 1989). Numerous studies illustratecognitive dysfunction, for example in learning and memorytasks (Beatty and Gange 1977, Beatty et al. 1988; Minden etal. 1990). In cognition as in motor and sensory abilities,dysfunction can be severe but is by no means evident in allpatients. There is considerable inter-patient variabilityeven at a given level of clinical disease. Rao et al.(1984) performed a cluster analysis to divide patients intogroups differing in memory performance, and commented: “Itis also surprising that the three subgroups, which appearedto differ sharply in their performances on measures ofmotor, cognitive, and personality functions, had equalaverage Kurtzke disability ratings”(p.631).Cognitive difficulties have even been detected inpatients with mild symptoms (Peyser et al. 1980b; Van denBurg et al. 1987), including patients with mild MS and inremission at the time of testing (Klonoff et al. 1991). Inthe view of Rao (1986) and Peyser and Becker (1984),cognitive dysfunction in MS has been underestimated becauseneurological examinations are not sensitive to subtle orspecific problems in cognition.9It is generally believed that, while mood/affectivechanges may be either organic or reactive, cognitive changesare usually organically-based (Peyser and Poser 1986),emerging directly from lesions. However, interaction canoccur between cognitive and affective domains, so that, forexample, mood changes can produce cognitive changes —depression can lead to decreased performance on cognitivetests (reviewed in Weingartner and Silberman 1982). Aswell, the interaction occurs both ways: cognitive changesmay influence a patient’s mood. For example, memorydysfunction may sufficiently frustrate some patients thatthey become short—tempered or depressed. To summarize,psychological changes may either have organic foundations orbe secondary reactions to the disease. Cognitive changesare generally considered organically—based, though they maybe secondary to affective problems; affective changes may beeither organic or may be responses to alterations incognitive or physical state. As well, psychological change,either organic or reactive, may be a feature even of earlystages of the disease.Psychological Profiles in MSTo examine the basis — reactive vs. organic — ofpsychological changes in MS, the assumption of this projectwas that particular changes do not necessarily occur in allpatients or in stages reflecting progression of the disease.10As will be outlined, numerous studies have tried todistinguish between reactive and organic psychologicalchanges in MS, often with equivocal results. Generally suchstudies have assumed that all patients show some degree ofabnormality on a specific dimension (e.g., depression). Incontrast, the current study takes the view that MS patientsmay demonstrate psychological change in various ways.Specifically, any two patients, even with similar levels ofphysical disability, may show quite different psychologicalprofiles. Based on the literature, it was postulated thatthese profiles may include:1. Depressive response, when a patient reportspsychological problems seemingly out of proportion tohis/her level of physical difficulty. Such patientscorrectly assess their physical status, so that subjective(patient report) and external (neurologist report)assessments of disability/impairment are concordant. Becausethis response is not assessed on the same basis asdepression in the psychiatric Diagnostic and StatisticalManual of Mental Disorders (DSM-IIIR) (though the twodefinitions of depression have numerous features in common),the response of this study will be referred to as‘Depression’.2. Denial, when a patient reports few problems eitherphysical or psychological, despite a relatively high scoreon an external assessment of disability/impairment.113. Exaggerated Somatic response, when a patientmagnifies physical symptoms in comparison to externalassessment, yet may claim that such symptoms have nonegative repercussions on mood or psychological well—being.4. severity—related response, when there is goodconcordance between a patient’s self—reported problems, bothmental and physical, and physical disability as assessed bya neurologist. This concordance should hold whetherproblems are few or many. In particular, it is notunrealistic to expect that a patient with severe physicalimpairment will be severely distressed.Any of the profiles may be purely reactive - asecondary effect of the illness, possibly representing anunwillingness or inability of the patient to cope with thediagnosis and/or the physical problems of MS.Alternatively, the profiles, notably ‘Depression’, Denial orExaggerated Somatic, may result from an organic processaffecting comprehension, mood, judgment, and insight.Therefore, two separate models may explain a reaction suchas, for example, denial. In the first model, the responseis reactive in a classic psychological sense; clearly somelevel of denial is frequently used by healthy individuals,and is considered part of normal behaviour (Freedman et al.1976). In the second model, the response is the result ofthe underlying organic process. For example, euphoria(associated with denial) and lack of insight have beenobserved in patients with frontal lobe lesions12(Filskov et al. 1981; Walsh 1987). These two models mayboth be represented to varying degrees in the patientpopulation, in that some patients who deny their illness areexhibiting purely reactive behaviour, while others who denyare manifesting organic pathology.Hypotheses Underlying the StudyThis study operated with four working hypotheses.1. Hypothesis one was that distinctive psychologicalprofiles of accommodation such as ‘Depression’, Denial,Exaggerated Somatic and Severity—related, do exist, even inpatients with similar levels of disease. The method bywhich profiles were detected, and by which patients wereseparated into psychological groups, will be describedshortly.2. Hypothesis two, based on support of the foregoingfirst hypothesis, was that membership in psychologicalgroups was not related to number of sites with lesions. Inother words, psychological abnormality or profile is not afunction of amount of biological MS. This hypothesisextended the central notion of this study, thatpsychological problems may not be stages in a diseaseprocess or reflective of amount of pathology, but rather mayresult from specific lesions.3. Hypothesis three was that membership in groups wasrelated to a different pathological indicator - location of13lesions. For example, patients with virtually no MSdisability but severe distress, or patients with extremephysical disability but apparent cheerfulness or evenelation, may possess brain lesions which compromise theintegrity of the central nervous system, and are in partresponsible for their psychological profiles. Membership incertain profile groups, then, may be related to, and aprimary result of, lesions detected by MRI. It wassuggested that in any particular profile either all members,or a subgroup thereof, may possess characteristic lesions.4. Hypothesis four was that group membership would berelated to performance on cognitive tests. These testsserve as an added measure (along with MRI) of organicpathology. It was suggested that patients’ coping stylesmay be a primary result of diminished cognitive capacity.By relating group membership (reflecting coping style) withboth MRI lesion-location and cognitive-function, two testswere performed for a potential organic basis to theseprofiles.If neither the third nor the fourth hypothesis wassupported, these data would suggest that coping strategiesare psychological overlays and reactions to the diseaserather than primary results of the neuropathologicalprocess.The purpose of this study, then, was to seek andidentify psychological profiles and organic correlates,through the testing of the above four hypotheses.14General MethodologyTo test for the existence of profile groups, it wasnecessary to assess each patient’s (1) actualdisability/impairment, (2) self-reporteddisability/impairment, and (3) psychological well-being withan instrument not confounded by MS signs and symptoms.Therefore, in the proposed study, two scales were derived:i) Sx: to assess patients’ own perceptions of their MS—related signs and symptoms, and ii) Ds: to assess patients’self—reported psychological well—being, or conversely levelof distress, independent of Sx. Psychometric evaluation ofthese scales was done to determine their reliability and, inpart, validity. In addition, the Expanded Disability StatusScale (EDSS) (Kurtzke 1983), which will be called the Kscale in this study, was used as an objective or externalmeasure of operational disability. The EDSS is a widely-used method of evaluating disability/impairment in MS. Apatient’s EDSS score can range from 0 to 10, and is based onscores in individual Functional Systems (pyramidal,cerebellar, brain stem, sensory, bowel and bladder, visual,cerebral/mental, and other/miscellaneous.)MMPI item data from 99 MS patients and 56 matchedcontrols were used to derive the Sx and Ds scales.The hypothesized profiles, operationally defined on thethree variables just described, were then conceptualized asillustrated in figure I. It was then possible to determine15FIGURE IPredicted psychological profilesConceptualization of profiles identified from the literature, as operationally defined on threevariables: (1) K (EDSS): neurologist’s assessment of the patient’s level of disability!impairment; (2) Sx: patient’s self-report of disability/impairment; (3) Ds: patient’s self-reportof psychological problems.lessdisabilitySx K DsExag. Somatic SeverityI I I I ImoredisabilityDenial‘Depression’JI I I I I16whether the hypothesized profiles of accommodation to thedisease occur in MS patients. This was accomplished byWard’s method of cluster analysis (Ward 1963) of MS patients(n=99) based on similarity of scores on the three scales(Sx, Ds, and K). The choice of the appropriate mathematicalsolution (number of clusters) was made using statisticalcriteria, namely maximization of between—group multivariatevariability and minimization of within-group variability fora manageably small number of clusters. The overall accuracyof the solution in terms of group membership was examined toestimate within-group homogeneity. These profiles were thenplotted and clusters with similar profiles combined.Subsequently, MRI data on these profiles were examinedto determine whether there was a relation between lesionlocation, or alternatively the number of sites with lesions,and group membership. The groups were also compared ontests of cognition which were appropriate by their lack ofsignificant motor and sensory components.As these data were derived from an ongoing researchproject, the positive results were validated on anadditional group of MS subjects (n=43).17Potential Contributions of the StudyThe work just described was designed to support orrefute the hypotheses that characteristic and distinctcoping responses exist in MS patients and that suchresponses have correlates in anatomical and/or cognitiveabnormalities. The study, and even a refutation of any ofits hypotheses, was designed to elicit information forresearchers, clinicians, and individuals affected by MS. Itmay be useful for patients, families, and health careworkers to know when psychological changes appear to be adirect result of the disease process. For the scientificcommunity, any knowledge regarding a relation between whitematter lesions and psychological responses will add tocurrent understanding of functional neuroanatomy.This project contributes to the development of whitematter models of brain function, through its analysis of thebehavioural consequences of disrupted connectivity in thebrain. Data from the project also add to the ongoing debatein neuroscience regarding localization of brain function.Despite the acceptance of anatomical localization for somebehaviours — few would disagree that body movement islargely mediated in anatomically discrete regions - thelocalization of some higher functions is still a topic ofdebate. The current project assumes the existence offunctional systems, to the degree that lesions in certainareas can at least influence complex emotional behaviours.18The results from the project are intriguing considering thegeneral lack of knowledge regarding the behaviourallyfunctional role of white matter regions.19CHAPTER TWOBACKGROUND TO THE STUDYOrganisation of the ChapterThis chapter presents a deeper foundation for theproject, outlined in Chapter One, by tracing the evolvingbody of knowledge on psychological aspects of MS. Majorstudies of psychological abnormalities in MS, both cognitiveand affective, are reviewed, with the aim of highlightingobservations of distinct psychological responses MS patientshave shown to their disease. The chapter summarizeshistorical debates on whether psychological disturbancescontribute to or even cause MS (“MS personality” theories)or, alternatively, result from the disease. With respect tothis latter position, the evidence for psychologicalproblems as either biological/organic orsituational/reactive, will be examined. Past correlationalstudies which have tested this distinction will then bediscussed. As these studies have generally reported weakcorrelations, it will be argued that these studies do nottake into account potential distinct responses (e.g.,depression vs. denial). Based on this progression of ideas,the conceptual foundations for the study will be outlined.Finally, methods to be employed will be explained.20For this review of the literature on psychologicalissues in MS, it should be noted that a distinction betweencognitive and affective domains was not made in early work,so that the two were often considered one. Before 1900, awide range of behavioural functions from memory to mood weresubsumed under the general term “mental” (Gowers 1896). In1922, Wechsler was discussing “mental” and “psychic” changesin MS as roughly analogous to biological and psychologicalchange. Sugar and Nadell (1943) called all changes, boththought and mood, “mental.” More recently, Trixnble andGrant (1982) use the term “psychiatric” to refer to, forexample, personality change, depression, and psychosis, and“neuropsychologic” to refer to cognitive disorders orgeneral dementia. During a historical discussion, then, asharp distinction between cognitive and affective is notalways possible. The current distinctions also havelimitations: there is considerable overlap between theconstructs of thought, mood, and personality. As always,categorization may clarify complex situations, but can alsooversimplify issues, and result in a loss of information.Psychological Aspects of MS: Several Distinct ResponsesMultiple sclerosis is a disease involvingneuropathological changes and their effects on motor andsensory systems. But since the earliest descriptions of the21disease, references have also been made to its psychologicaland psychiatric aspects.One of the best—known of early anecdotal descriptionsconcerns the case of Augustus d’Este (1794—1848) whoseletters and diaries suggest that he suffered from MS forabout half of his 55 years. Several of his diary entrieshave been interpreted as evidence of mental symptoms of MS(Stenager et al. l989a), and a reading of the text (Firth1948) reveals at least one stress—related attack.Inconclusive though the report may be, it hints thatpsychological factors have always been entwined with MS.Medical descriptions of MS began to be published in the19th Century, most notably through the writings and lecturesof neurologist Jean Martin Charcot. By the 20th Century, MSwas an accepted clinical entity. However, cases of whatwere probably multiple sclerosis were recorded considerablyearlier. For example, the earliest known descriptioncompatible with the diagnosis of MS (Hashimoto and Paty1986) was that of Lidwina of Schiedam (1380—1433) (Medaer1979)Nineteenth-century descriptions indicated thatpsychological problems were part of the disease, though suchdescriptions are difficult to appraise in twentieth-centuryterms because of changes in concepts of mental health in theintervening 100 years. For example, Seguin (1878)pronounced one male MS patient psychologically sound partlybecause he did not masturbate. As pointed out by Aring22(1965), the history of clinical assessments shows that moraljudgments have frequently been intrinsic to diagnosis,notably in the zealous application of the diagnosis ofhysteria. Sequin (1878) reported that a female patient withweakness and paresis showed inappropriate affect (“at timeshysterical laughter and tears”), and “concluded that thepatient had a functional palsy of a hysterical nature,” onlyto discover at autopsy that she had numerous scleroticlesions. Buzzard (1897) also described nine cases of MSwhich had mistakenly been diagnosed as hysteria. Thisdiagnostic dilemma is more than a historical curiosity. Forexample, Skegg et al. (1988) reported that 15/91 (16%) of agroup of MS patients were given psychiatric diagnoses,including hysterical conversion and hysterical personality,before their MS was diagnosed. The term “hysteria” applieswhen patients have physical problems, but the cause of whichappears to be psychological rather than biological.Examples are the somatoform disorders of DSM-IIIR, includingconversion disorder in which patients have paralysis,anaesthesia, aphonia or any of a number of other physicalmanifestations of what is judged to be psychologicalconflict. However, some observers have commented that theterm hysteria has been applied carelessly (Brown and Davis1922) and as a value judgment of the patient (Aring 1965).Certainly, hysteria has been “an imprecise term” (Trimbleand Grant 1982).23Probably the most influential observer of early MS wasCharcot (1877), whose graphic account of psychologicalabnormalities has echoed through the MS literature eversince. Said Charcot, MS patients showed “markedenfeeblement of the memory. . . [and an] almost stupidindifference in reference to all things. . . . It is notrare to see them give way to foolish laughter for no causeand sometimes, on the contrary, melt into tears withoutreason” (p. 194)From the turn of the century to the l920s, MS reportsconsisted generally of case studies (Dercum 1912; Brown andDavis 1922) noting a variety of mental symptoms. Most suchstudies were on a small number of pre—selected patients. Inthe main, studies from this period suffered from a lack ofobjective psychological measures and overly ambitiousgeneralization from the data. Jelliffe (1921) inferred onthe basis of two patients’ dreams that MS may have beencaused by “illness in the spiritual part” (p.675) of theirnatures. It is instructive when evaluating such work torecall how much less was known about brain function early inthe century, and that pure speculation was much moreacceptable in both journals and books than is the casetoday. It was in this period that the popular Common SenseMedical Adviser (Pierce 1909) informed readers that“excessive intellectual activity” (p. 125) was liable tocause brain damage and serious illness. “The production of24thought wears away the gray matter of the cerebrum” (p.124)As can be seen from the above references, early work onthe psychology of MS drew attention to the possibility thatpsychological changes were both frequent and significant inMS. An examination of these early studies also reveals thatthey recognized another phenomenon: the existence of severaldistinct abnormalities or responses in MS patients. Earlystudies reported that some MS patients showed a tendency tofabricate or exaggerate physical problems (Brain 1930),while others showed “euphoria” (Cottrell and Wilson 1926),and yet others showed depression (Brown and Davis 1922). Inrecent years, the influence of stage—theory in psychologyhas prompted the suggestion that MS patients may go througha series of emotional stages after diagnosis (Matson andBrooks 1977). However, distinct abnormalities havecontinued to be observed and reported even as research hasbecome more large—scale and improved measures have beenused. As will be seen, the abnormalities found have notcorrelated well with disease variables, suggesting they arenot necessarily a function of degree/amount of disease.A report based on 1,970 case records, published soonafter World War I (Wechsler 1922), concluded that MS was acommon disease in the United States, and that it wasfrequently accompanied by “psychic” symptoms such asirritability, depression and general nervousness, and by“mental” ones such as subtle dementias. Wechsler commented25that numerous psychological problems were distinctive butnot characteristic, in that not all patients exhibited them.One large-scale study of psychological changes in MSwas completed by Borberg and Zahle (1946). They summarizeddata from 330 MS patients, each of whom had been examined byat least five different physicians for psychologicalabnormality. Borberg and Zahle did not report the criteriaused for assessment of abnormality, but did report that 47%overall developed “mental symptoms” of various kindsincluding euphoria, depression and dementia.Paradoxically—elevated mood has also been reported inMS patients. Dercum (1912) cites the inappropriate laughterobserved in the disease, as had Charcot (1877). Influentialin thjs regard was a 1926 paper by Cottrell and Wilson,based on a list of 48 questions on mood and emotion,administered to 100 patients, in which the authors concludedthat emotional change was common and serious in MS.Commenting that every single patient exhibited some moodchange, Cottrell and Wilson reported that euphoria, aninappropriate sense of mental well—being, was exhibited by63% of patients, and that it was associated with denial ofillness. Another affective state was an inappropriate senseof physical well-being in the form of a lack of recognitionof physical disability. The authors called this state“eutonia” (p.8), and said it was shown by 84% of patients.A separate investigation in the same period (Ombredane 1929)added to this work, suggesting that cognitive problems26existed in some patients as did affective problems inothers. In that study, more than 70% of 50 patients hadcognitive problems, virtually the same percentage aspatients who showed affective problems.Methodologies varied widely, and at times were notreported, earlier in this century, which contributed todiffering prevalence rates from study to study. However,the general conclusions were that (1) psychological changesdo accompany MS, and (2) several distinctly differentchanges can be observed. Some of the disagreements throughthe years about the prevalence of psychological dysfunctionin MS stemmed from the fact that brief mental statusexaminations were frequently used in neurological studies,while detailed neuropsychological tests were more often usedin psychological studies. Generally, studies bypsychologists have shown more mental involvement in MS thanstudies by neurologists (Stenager et al. l989a; Peyser etal. 1990). For example, Peyser et al. (l980b) found that MSpatients who were judged mentally intact on neurologicalexamination frequently showed cognitive dysfunction on moredetailed neuropsychological tests. Brief mental statusexaminations may be inadequate for the detection of subtleabnormalities (Peyser and Becker 1984; Mahier et al. 1989).However, neurological studies have made contributions whichlarge—scale neuropsychological studies have not, includingin the elegant descriptions of individual patients’presentations.27The 1940s and 1950s saw the emergence of psychometricsand the development of more objective tests of mental andpsychological function. MS was now studied usinginstruments such as the Minnesota Multiphasic PersonalityInventory (MMPI) (Canter l95la; Baldwin 1952); WechslerBellevue intelligence test (Canter 1951b); and the Halsteadbattery to assess “biological intelligence” (Halstead 1947)(the forerunner to the Halstead-Reitan test battery usedtoday).In cognitive function, as in affective function,numerous studies suggested that problems exist, but that notall patients display them. Canter (1951b) was one of thefirst to establish conclusively that MS can affectcognition; he showed declines in the cognitive test scoresof American armed forces veterans who developed MS, over aperiod of just six months. Canter also pointed out thelarge standard deviations seen in patients’ test scores,reflecting considerable variability, and thereforesupporting the notion of distinct responses. A large numberof more recent studies have also shown that MS patients canexperience cognitive dysfunction. Often cited have beendeficits in memory (Jambor 1969; Beatty and Gange 1977; Raoet al. 1984; Beatty et al. 1988; Litvan et al. 1988). Raoet al. (1984) showed that chronic progressive MS patientscan experience deficits in both verbal and visual—spatiallearning tasks. Numerous studies have also made it clearthat cognitive problems are not necessarily late—stage28effects only, but can occur early in the disease (Peyser etal. 1980b; Grant et al. 1984; van den Burg et al. 1987;Klonoff et al. 1991). As Rao (1986) has said in aliterature review of the neuropsychology of MS, many suchstudies show large standard deviations in cognitive testdata. “This observation implies that some MS patients maydemonstrate little, if any, cognitive dysfunction, whileothers may exhibit moderate to severe disturbance” (p. 530).Conceptual and abstract reasoning has also been citedas a problem in MS by numerous investigators of the lastseveral decades (Rao 1986). Tests of abstract reasoningshow that MS patients perform worse than non—brain—damagedcontrols (Peyser et al. 1980b; Rao et al. 1984) and that MSpatients’ performance is similar to that of brain—damagedcontrols (Matthews et al. 1970; Goldstein and Shelly 1974).In the affective domain as well as the cognitive,increasingly detailed reports are being published on theprevalence and severity of dysfunction in MS. Depression isprominent (Baldwin 1952; Whitlock and Siskind 1980; Schifferet al. 1983; Joffe et al. 1989). Studies of the prevalenceof significant depression have reported a range of 27-54%(Pratt 1951; Surridge 1969; Whitlock and Siskind 1980;Schiffer 1983; Joffe et al. 1989). Again, not all patientsshow the problem.Euphoria has also long been reported and continues tobe so, though there is no consensus on its prevalence, withestimate rates ranging from 0% (Baldwin 1952) to 63%29(Cottrell and Wilson 1926). Differences in observeddysfunction result partly from a continuing lack of reliableand standardized methods for diagnosing euphoria (Minden andSchiffer 1990) but may also result from sampling bias. Itmay be that only a subgroup of MS patients show euphoria,and that the idiosyncratic selection criteria of differentstudies result in variable sampling of this subgroup.Response Types Identified in the LiteratureDepressionDepression in MS has been noted repeatedly in theliterature (Baldwin 1952; Goodstein and Ferrell 1977;Whitlock and Siskind 1980). While some MS patients aredepressed, others are not, so the population is“psychiatrically heterogeneous” (Jambor 1969, p. 767). Aswell, numerous reports have noted that some MS patients showa depression out of proportion to their apparent physicaldisability or impairment. Peyser et al. (1980a) did acluster analysis of 52 MS patients using a variety ofphysical, cognitive and psychological measures, and showedthe existence of a group whose members were acutelydistressed despite a relative lack of physical or cognitiveproblems. Zeldow and Pavlou (1988) also did a clusteranalysis of 81 MS patients and identified a group whosemembers were characterized by unhappiness and distress30regarding their illness, although they were not any morephysically impaired than other patients. Other authors havereported depression preceding the onset of more—typical MSsymptoms (Young et al. 1976; Goodstein and Ferrell 1977);such depression is therefore also non—concordant with MSsigns and symptoms.DenialPeyser et al. (l980a) reported that 13 members (25%) ofa cluster—analysed sample showed denial of distress ordifficulty. Zeldow and Pavlou (1988) performed a clusteranalysis of MS patients using factors from the CaliforniaPsychological Inventory, and found that a group of 18 (22%)was “unusually concerned with creating a favorableimpression and with denying any worries or difficulties”(p.193). Denial among MS patients has also been observed byothers (Gilberstacit and Farkas 1961; Surridge 1969). MSpatients have also been described who show lack of expectedanxiety or a seemingly inappropriate sense of physical wellbeing. Cottrell and Wilson (1926) cited patients who saidthey felt physically well when they were in fact disabled.According to this report, such “eutonia” was a commonphenomenon and was shown by 84% of the 100 MS patientsinterviewed. MS patients have also been described who showa lack of expected anxiety. Many of these patients havebeen called euphoric. Though an imprecise term, euphoria is31related to the phenomenon of denial (Surridge 1969;Weinstein 1970). Many authors have observed euphoria in MS,which Cottrell and Wilson (1926) defined as an inappropriatemood of cheer, happiness and ease despite physical problemswhich would be expected to elicit tension or anxiety.Recent reports have expressed caution about vague labelswhich confuse prevalence assessments, but generally agreethat euphoria exists (Baretz and Stephenson 1981; Rabins etal. 1986).It should be noted that the concept of denial caninclude either or both of: (1) a lack of recognition, ordisavowal of, reality; (2) a recognition of reality, but anapparently abnormal lack of anxiety in the face of thatreality (Strauss et al. 1990). Denial is sufficientlycommon, and at times sufficiently maladaptive, that Strausset al. (1990) have suggested that the next revision of theDiagnostic and Statistical Manual of Mental Disorders (DSMIV) of the American Psychiatric Association should include anew subtype of adjustment disorder called “maladaptivedenial of physical disorder” (p.1168).Exaggerated SomaticNumerous authors have suggested that MS patientssometimes exaggerate problems and create non—existing ones(Brain 1930; Weinstein 1970). “Hysterical symptoms, such aspareses and ataxia, seem to occur more often in association32with disseminated sclerosis than with any other organicdisease of the nervous system” (Brain 1930, P. 372).Though a distinction must be made between hysteria, andexaggeration of somatic complaints, the two havecharacteristics in common. Hysterical neurological problemsare physical realities, for example paralysis or sensoryloss, the roots of which are psychological. Exaggeration ofsomatic complaints involves extreme concern despite acomplete or relative absence of physical problems.The concept of hysteria has been applied in MS sincethe disease was first described. The label has been looselyused for different groups: (1) patients who are recognizedto have physical problems but who are believed to beexaggerating those problems, and (2) patients whoserelapsing sensory and motor disabilities are thought to bepsychogenic. Skegg et al. (1988) showed that numerousindividuals with MS are first given diagnoses of hysteria.Peyser et al. (l980a) showed some patients’ tendency tofocus on, or exaggerate, physical problems, by demonstratingthe existence of several subgroups of patients with anunusual degree of somatic concern. One such subgroup,constituting 13/52 (25%) of the sample, displayed a concernout of proportion to their moderate impairment. “Suchpatients will probably be in the physician’s officefrequently and require as much emotional support andreassurance as medical assistance” (pp. 438—439). Anothersmaller subgroup, 3/52 (5.8%), showed minimal physical33disability but “hysterical personality style” in the form ofelevated MMPI scores on Hysteria and Hypochondriasis. Theauthors commented that, in the case of such patients,physicians may always have some doubt about the accuracy ofthe MS diagnosis, and should consider the possibility ofconversion reactions occurring alone or during exacerbations(Peyser et al. 1980a).Severity-relatedSome studies point out that patients can react inapparent concordance with their physical problems. That is,those with mild disability show mild distress, and thosewith more disability show more distress. Baretz andStephenson (1981) report that 4/40 (10%) of MS patientsinterviewed had a “realistic recognition”(p.119) of theirlimitations and physical problems. Peyser et al. (1980a)also defined a cluster of patients whose profound distressmatched their severe physical problems.34Psychological Problems:Cause or Result of the Disease?Numerous reports through the middle decades of thetwentieth century interpreted data in the light ofpsychodynamic or psychosomatic theories of disease. Ineffect, such reports suggested that the psychologicalproblems of MS are a cause, rather than a result, of thedisease. Inman (1948) reported, based on a small number ofpatient interviews regarding attitudes toward parents andsexuality, that MS was a somatic response to intolerablemental conflict. Philippopoulos et al. (1958) took anunspecified detailed history of each patient after which thepatients’ personality configurations were formulated inpsychodynamic terms. He reported that MS patients werevulnerable to a specific dynamic process based on emotionaland psychosexual immaturity due to early frustrations, whichmakes them susceptible to “psychosomatic disintegration” (p.472). Grinker et al. (1950) interviewed 26 MS patients,employing the Rorschach ink blot on a handful, and concludedthat MS patients were unusually immature and possessexcessive needs which their human relationships could notsatisfy. Such studies were interpreted as supporting thenotion that MS patients have particular personalities whichcause or exacerbate the disease. Subsequent authors (Rikianet al. 1961; Vander Plate 1984; Peyser and Poser 1986) havecautioned against hasty interpretations from such data.35Some authors have said that early investigators too oftenmade “broad generalizations and conclusions” (Rikian et al.1961) from single tests, or used “anecdotal” evidence(Surridge 1969). A study by Pratt (1951) providedalternative evidence to the controversial MS—personalityhypothesis. Pratt assessed 100 MS patients, compared with100 controls, for similarities in personality by allottingpatients to classes after the schemes of Jung (extraversion—introversion) and Sheldon (viscerotonia, somatotonia,cerebrotonia). He found that MS patients did not appear tobe of any one personality type, nor did the numbers in eachpersonality class differ significantly between MS andcontrols. There were also no differences between the twogroups in hysterical manifestations, obsessional traits,psychopathy, childhood environment or early separation fromparents.In recent years, support has shifted away from thenotion of a typical premorbid MS personality (Vander Plate1984). However, personality factors may not be completelyindependent from the disease. There is some evidence thatstress may affect MS (Mei—Tal et al. 1970; Grant 1985), andwhether a particular event is stressful is partly subjectiveand personality—based. Nevertheless, many researchers wouldnow say that psychological problems are most likely a resultof the MS disease process itself in some way (Vander Plate1984; Peyser and Poser 1986).36Psychological Problems as a Result of the Disease:Reactive or Organic?If psychological problems result from the disease, thequestion remains whether the problems result directly fromthe lesions, or are reactive responses based on theindividual personality involved. Such a question appliesparticularly to behaviours such as depression and denial,which can be normal and which may exist without braindysfunction. For example, mourning after a loss isexpected, and only becomes clinical depression if prolonged.To the question of whether changes are reactive ororganic, support has been marshalled for both sides. On oneside, studies including the following have provided supportfor the notion of a reactive basis to psychologicalproblems. Surridge (1969) showed that depression in MS issimilar to that in muscular dystrophy, which is also chronicand disabling but which does not involve the central nervoussystem. Logsdail et al. (1988) showed that there is asignificant correlation between the severity of psychiatricsymptoms and patients’ degree of social stress, but notbetween severity of psychiatric symptoms and the presence ofMRI abnormalities. Jouvent et al. (1989) showed thatrecent—onset MS patients are more likely to show depressionthan are patients with long-standing disease.On the other side, studies including the followingsupport the notion of an organic basis to psychological37problems. Schiffer et al. (1983) showed that there is moremajor depression in patients with clinically-inferredcerebral involvement than in patients who do not have suchinvolvement, but rather who have (clinically—inferred)spinal cord and cerebellar lesions. According to Whitlockand Siskind (1980) and Joffe et al. (1989), MS patientssometimes experience serious depressive episodes months oryears before physical symptoms appear. Rabins et al. (1986)showed that MS patients with brain involvement (assessed byCT scan) were more likely to be identified as euphoric thanwere patients with only spinal cord involvement. AndBraceland and Giff in (1950), Surridge (1969), and Rabins etal. (1986) showed that euphoria is associated with cognitivedeterioration. Then, numerous studies can be interpretedeither way. For example, Dabs et al. (1983) showed that MSpatients with progressive, non—remitting disease show moredepression than patients with relapsing-remitting MS.Many correlational studies have been done in MS,involving the examination of pairs of phenomena to directlyor indirectly address the question of whether psychologicalabnormalities are reactive or organic. As can be seen froman overview of these studies (Appendix A) data have notconverged, and clear conclusions have not been reached.Where correlations have been found, these have often beendisappointingly low. Some of the studies have producedseemingly paradoxical results, such as that of Logsdail etal. (1988) which reported that psychiatric problems were38more severe in patients without MRI—imaged lesions than inpatients with such lesions. However, those patients withmore severe psychiatric problems also scored high onmeasures of social stress (related to work, finances,housing, and social, marital and family circumstances);therefore, degree of social stress was a better predictor ofpsychiatric problems than were MRI lesions, emphasizing thecomplexity of psychological change in MS. The importance ofsocial factors in some cases of affective disorder wasrecognized by Schiffer (1987), who commented on theheterogeneous nature of depression in MS, and cited fourseparate categories of the disorder (biological, social,psychological and other).As can be seen, a reading of the literature revealsapparent contradictions. While individual correlationalstudies can provide valuable information, lack ofconsistency in methodology make discrepancies in theliterature difficult to analyse (Peyser et al. 1990; Mindenand Schiffer 1990). Methodological differences includeresearchers’ choice of (1) subjects’ disease course, (2)measure of extent of either clinical or biological disease,(3) measure of mood or cognitive abnormality, and (4) methodof assessing and scoring MRI lesions. There are numerousother methodological differences as well. For example, insome studies all variables are continuous, while in otherstudies one or more variables are categorical or ordinal.39However, even if methodologies were uniform, linearcorrelational studies may not reveal strong relationsbetween psychological variables and pathological ones,because such studies generally do not take into account theexistence of qualitatively different psychologicalabnormalities.The Study: ConceptuallyThis study relied on a progression of ideas which willbe summarized here. MS patients do show psychologicalresponses to their illness. As well, the evidence suggeststhat several qualitatively distinct responses occur, whichwill be referred to as psychological profiles. When twopatients show different profiles, the difference may be afunction of lesion—location or personality—based behaviouraltendencies rather than of disease stage. Therefore,identification and separation of patients into differentprofile—groups allowed for clearer analysis of the aetiologyof psychological responses. Profiles identified from theliterature have been described earlier. The current studyfirst operationally defined such profiles, and then by meansof cluster analysis determined whether such profiles existedin a sample of MS patients. To define such profiles, it wasnecessary to have measures of objective disability, self—report disability, and mood/distress. Individuals with a‘Depressed’ profile would be expected to show significant40distress despite little disability, either objective orself-reported. Those with a Denial profile would have beenexpected to report few problems, either physical or mood—related, although an objective assessment would have foundconsiderable physical problems. Those with an ExaggeratedSomatic profile would have had a much higher level of self-reported physical disability than disability objectivelyassessed; these individuals would not necessarily have beendistressed. Others with a Severity-related profile wouldhave scored similarly on all three measures, either low orhigh, so that self-reported disability was concordant withobjective report of disability, with distress onlyreflecting amount of disability. Three variables weretherefore assessed (amount of distress, level of physicaldisability, and perceived level of physical disability)which, taken together, defined a profile.The current study therefore avoided the pitfall ofassuming that all MS patients show some level of, e.g.,depression, when many patients may not. As well, thisstudy’s multivariate approach allowed for separation, notonly of distressed patients from non—distressed ones: itwent one step beyond, to allow separation of those who aredistressed and who have many physical problems, from thosewho are distressed despite few physical problems.Cluster analysis was used because it is a statisticaltechnique for determining whether elements of a large groupnaturally fall into distinct subgroups on given measures.41Cluster analysis has been used in several MS neuropsychologystudies (Peyser et al. 1980a; Rao et al. 1984; Rao et al.1989a; Fischer 1989), and in an MS personality study (Zeldowand Pavlou 1988) but never in the way here proposed. Rao etal. (1984) and Fischer (1989) clustered patients based onmemory performance; Rao et al. (1989a) clustered patientsinto two cognitive groups, one with relatively substantialimpairment and one with relatively minimal impairment.Peyser et al. (1980a) grouped patients on a variety ofmeasures, including cognitive and affective, and identifiedseveral distinct responses. Zeldow and Pavlou (1988)grouped patients on four personality—related factors of theCalifornia Psychological Inventory, and identified distinctgroups. Unlike the above studies, however, the currentstudy stated that there may exist certain profiles whichcould be operationally defined by measures chosen to capturecertain domains of interest. The three measures formed thecore definition of the profiles.The study, then, was designed to establish whether MSpatients are truly psychologically heterogeneous in responseto illness, which would have strong implications forneuropsychological research in MS. It employed a simple butinnovative method for separating different profiles one fromanother, creating the possibility of stronger correlationsbetween such profiles, and lesion data.The second objective of the study was to determinewhether membership in a profile group (e.g., Denial) was42related to the site of MS lesions. MRI data were thereforeassessed to see whether members of particular profiles hadlesions in common. If such relations were found, one wouldthen have evidence supporting an organic basis to theprofile.An analysis of lesion data was based on the recognitionof MRI as the best available technique for visualization ofMS lesions disseminated in space (Paty et al. 1988). TheMRI image is realized by the emission of electromagneticradiation from the nuclei of hydrogen atoms after they havebeen excited by radiofrequency pulses in a constant magneticfield. Because water is a major component of human tissue,and because every water molecule contains two atoms ofhydrogen, the MRI image is largely based on water, both onits quantity and on its macromolecular environment. Sincedifferent tissues vary in their water content, and edematousand deinyelinated regions contain more water than do normallymyelinated tissues, NRI can distinguish lesioned areas fromnon—lesioned ones. The unparalleled tissue contrast of MRIhas made it very useful for MS clinicians and researchers.Several limitations were, however, taken into accountin analysis of the MRI data. One was that MRI is notperfect in either sensitivity or specificity. Regardingsensitivity, even in patients with clinically definite MS(CDMS) (Poser et al. 1983) MRI does not always showabnormalities in 100% of such patients. Paty et al. (1988)report abnormal MRI scans in 93% of CDMS patients. One43 -possible reason is that MRI scans are usually done of thehead, not the spinal cord, yet spinal cord lesions maycontribute to clinical signs and to a diagnosis of CDMS.Another possible reason is that lesions in the head mayimpair behavior but be too small to be detected by MRI.Regarding specificity, MRI sometimes showsabnormalities in subjects with no known MS (Paty et al.1988). Such abnormalities occur, for example, inindividuals with cerebrovascular disease (Ormerod et al.1984; Gerard and Weisberg 1986). MRI-detected lesions alsoappear in normal controls (Ormerod et al. 1987; Logsdail etal. 1988; Hunt et al. 1989), particularly with age and inperiventricular areas. The specificity of MRI to MS wasinvestigated by Yetkin et al. (1991) who examined scans from92 MS patients, and 168 other subjects who had hypertension,dementia, or no known illness. Specificity (the proportionof non—MS subjects whose images were correctly classified asnon-MS) was 95%-99%, indicating a small risk thatperiventricular white matter abnormalities can be wronglyinterpreted as MS (Yetkin et al. 1991).Lesions in normal controls have most often, but notexclusively, been reported in subjects over the age of 50.Hunt et al. (1989) reported that 20-30% of 46 normalsubjects over the age of 65 had white matter lesionsdetected by MRI. There have been several reports ofabnormalities in apparently normal individuals over 50(Gerard and Weisberg 1986; Fazekas et al. 1988; Kertesz et44al. 1988) and over 60 (Braffman et al. 1988). Reports oflesions in younger subjects have been few, although Harveyet al. (1990) reported small white matter hyperintensitiesin 7 out of 36 normal controls under the age of 50.Subjects in the current study are under the age of 50.Another MRI limitation was that each anatomical sitecontains many nuclei and tracts. This touches an importantissue for any localization study: In what manner should thebrain be divided into regions? Theoretically, there aremany different ways in which the brain could be divided. Itcould be divided based on a knowledge of neuroanatomy, sothat a functional system constitutes a region. At the otherextreme, it could be divided without reference to function,on a grid system. This study took a middle ground: areaswere chosen based on previous experience with MS lesions, sothat a known high—incidence anatomical area could be viewedas one region. In localization studies there is also thequestion of number of regions. Considering the complexityof the nervous system, countless regions could be specified.Because analysis of a large number of sites is impractical,how many sites should be delineated for methodologicalpracticality yet recognition of the intricacy of the nervoussystem? The 50 sites of this study do not amount to finedivision of the brain. Nevertheless, considering the lackof knowledge of the biological basis of psychological changein MS, division of the brain into 50 sites was considered agood point of departure.45The third objective of the study was to determinewhether membership in a group was related to the number ofsites with lesions. If no such relation was found, then amodel based on stage of biological disease would not besupported. It is important to note that number of siteswith lesions is a measure of biological disease activityrather than severity of clinical signs; there is not a highcorrelation between clinical and biological MS at least whenthe latter is measured by the extent of lesion areaidentified by MRI (Paty et al. 1985). It is also importantto note that there is no single measure of extent ofbiological disease in MS. The measure used in this study(number of sites with lesions) has the advantage ofencompassing a component of scatter or dissemination, whichmeasures such as total lesion area do not.The fourth objective was to determine if specificcognitive abnormalities were associated with groupmembership. A comparison of groups on neuropsychologicaltest performance was an additional method for determiningwhether members of any profile had cognitive dysfunctionsuggestive of an organic basis to their profile. Members ofthe Denial group, for example, might have been expected toshow dysfunction on cognitive tests requiring conceptformation and insight, such as the Halstead Category test orthe Similarities subtest of the WAIS-R (Lezak 1983).Members of the ‘Depression’ group may have had problems intests of new learning and memory (Nott and Fleminger 1975;46Kiloh 1961), such as Sentence Repetition, or PairedAssociate Learning (Lezak 1983).The Study: Rationale for MethodologySublectsMS subjects in the study were chosen because they had aclinically mild disease course. Data were therefore notconfounded by late-stage deterioration, either physical orpsychological, which might have interfered with testing.All patients had a diagnosis of relapsing-remitting MS; thisallowed the study to control for the cyclic nature of thedisease and search for subtle psychological change bytesting only patients in remission. All patients were freeof drugs (MS-related or otherwise) which might have hadpsychological effects, and no patient was included who had ahistory of psychiatric problems pre-dating the diagnosis ofMS: in that way, pre—morbid functioning was controlled.Patients were well—matched with normal controls on sex, ageand education, critical parameters of patientcharacterization which Peyser et al. (1990) note have beeninadequately matched in many past studies.47MeasuresThree measures were used for definition of thepsychological profiles. The first measure assessedpatients’ MS signs and symptoms, scored by a neurologist onthe Kurtzke Expanded Disability Status (EDSS) scale, themost widely-used rating system for MS physical impairment.The second measure assessed patients’ own perceived MS signsand symptoms, self-reported on the Minnesota MultiphasicPersonality Inventory (MMPI). A subset of MMPI questions,recognized as reflecting physical problems common in MS,form this measure. The third measure assessed patients’distress or psychological well-being independent of MS signsand symptoms, again self—reported on the MNPI. A subset ofMMPI questions reflecting mood form this measure.The two latter measures were developed as part of thisstudy. For measure (2), it has been noted previously thatthe MMPI contains items which are symptoms of MS (Baldwin1952; Marsh et al. 1982). This has been deemed problematicwhen the MNPI is administered to MS patients withoutrecognition of the inclusion of such items (Marsh et al.1982); however, their inclusion was usefully employed in theproposed study. For measure (3), the MMPI as a wide-rangingtest of emotional status also contains items assessingdistress and psychological well-being. Reliability analyseswere conducted for each of these two scales, using48Cronbach’s alpha (Cronbach 1951) as the internal consistencymeasure.Cluster AnalysisMethodologically, cluster analysis was appropriate forthe question of interest. Not only can cluster analysisidentify naturally—occurring groups in heterogeneouspopulations, but it requires no a priori assumption that theunderlying variable is normally distributed.Lesion AnalysisMRI was used for this study because it is superior toother imaging techniques in showing the demyelinated lesionsof MS. As well, because MRI is the imaging method of choicefor MS, there is a need for psychological data in relationto MRI. In particular, while there is some literature onthe effects of lesion burden in MS (Appendix A), there isless literature on the psychological effects ofanatomically—specific lesions.Neuropsychological TestsBecause neuropsychological tests can suggest thepresence of brain dysfunction (Lezak 1983), such tests canprovide a second indicator, along with MRI, of an organic49basis to psychological profiles. Tests were employed whichhave little or no motor or sensory component. Tables 1 and2 list relevant characteristics (Lezak 1983) of the testsemployed, which are by design relatively simple in terms ofcognitive dimension assessed, compared with broadintelligence tests. Table 1 outlines large—scale categoriesof cognition said to be measured. Some such abilities canbe a result either of nature (hereditary potential) ornurture (e.g., schooling), but this potential confound wasnot a factor in the proposed study considering that patientsand controls were closely matched on demographic variablesincluding education. Table 2 gives an overview of graymatter regions which, when damaged, result in decreases intest scores. This overview may serve as a generalreference, although it is not known to what degree whitematter lesions produce similar dysfunction.50TABLE 1Neuropsychological tests used:dimensions of cognition measured by each test(a) WAIS-RDigit Span attention; immediate memory capacityArithmetic attention; immediate memory; remote recallInformation verbal; remote memoryVocabulary verbalComprehension remote memory; verbal skills; judgementBlock Design visuo-spatial organisationObject Assembly visuo-spatial organisationPicture Completion visual acuity; visual organisationPicture Arrangement sequential thinking; visual spatial org.,social understandingSimilarities verbal concept formation; abstraction(b) non-WAIS-RDichot. Listening attention; immediate memory capacity;lateral preferenceWord Fluency remote memory; verbalSpeech Perception verbal recognition; phonetic abilityHaistead Category abstraction; concept formation; attention;immediate memoryPaired Assoc. Learn, verbal learning and memoryTrails B—A speed of visual information—processingSentence Repetition verbal memory; memory spanBenton Visual Memory visuo—constructive abilities; visual—spat ial perception/memory IMemory for Objects immediate visual memory51TABLE 2Neuropsychological tests used:gray matter lesions thought to be associated(a) WAIS-RDigit SpanArithmeticIn format iondom.hem.x (?)xxxxVocabularyComprehensionBlock DesignObject AssemblyPic.Completion xPic . ArrangementSimilarities(b) non-WAIS-Rnon-domhem.xxxxnon- domhem.xxfrontallobexxxfrontallobexxpar.lobexxxxpar.lobextemp.lobexxxxxtemp.lobexxxoccip.lobexxxxoccip.lobexxxDichot. Listen.Word FluencySpeech Percep.xxdon.hem.xxxxxxxxHaistead Cat.Paired A.Learn.Trails B-ASentence Rep.Benton Vis.Mem.Mem.for Objects52CHAPTER THREEMETHODSSublect Selection and CharacteristicsSubjects were drawn from an ongoing study, for whichparticipants were selected as follows. Potential MS-patientvolunteers were identified at their annual visit to the UBCMS Clinic, according to certain entry criteria:(1) a diagnosis of clinically definite MS (CDMS) asdefined by Poser et al. (1983) as follows:a) two attacks and clinical evidence of twoseparate lesions, or:b) two attacks, clinical evidence of onelesion, and paraclinical evidence ofanother, separate, lesion.(2) relapsing—remitting course and in clinicalremission at the time of assessment.(3) age less than 50.(4) age at onset less than 40.(5) ambulatory and functionally independent.(6) no other significant medical condition.(7) no history of psychiatric illness prior to thediagnosis of MS.(8) taking no medications at the time of assessment.(9) no prior neuropsychological examination.53Each MS subject was asked to identify a non-relatedcontrol who was as much like the MS subject as possible on alist of demographic variables: (i) age, (ii) sex, (iii)education, (iv) marital status, (v) occupation whenemployed.Controls, like MS subjects, were required in additionto satisfy these criteria: no history of significant medicalcondition including psychiatric illness; no current druguse, either prescription or non—prescription; and noprevious neuropsychological examinations.Data from 99 MS patients and 56 normal controls wereused in the initial study. Data from further subjects, bothMS and normal controls, were used in a subsequent validationstudy.Test/Assessment ProceduresPotential subjects, identified at their annual visit tothe UBC MS Clinic, received a standard neurologicalexamination, including assessment for physical disabilityaccording to Kurtzke Functional Scales and the EDSS (Kurtzke1983)Patients who agreed to participate in the study weregiven an appointment date within one month of their clinicvisit. On that test date, subjects were asked whether theirphysical conditions had changed noticably in the interim; ifso, they were re—assessed neurologically. Demographic data54on patients were collected during formal intake interviewsusing a structured format. The project was described indetail to patients, who then gave informed consent.The test procedure, which took place over several hourson one day, included a brain scan by magnetic resonanceimaging (MRI). MRI scans were performed on a PickerInternational Cryogenic 2000 MR scanner at a field strengthof 0.15 Tesla. Contiguous slices were obtained at 10 mmintervals in transverse and sagittal planes, using a double-echo spin echo pulse sequence with repetition time of 2,175msec, and echo delay time of 60 and 120 msec. Scans wereread by radiologist Dr. David Li, who recorded anyabnormalities greater than 2—3 mm in diameter.Patients were also given psychological tests,including the MMPI and a neuropsychological battery. Atrained psychometrician administered the tests. Because MSpatients can fatigue easily, patients were given occasionalbreaks if that appeared to be needed. So that fatigue didnot affect patients’ scores on particular tests which mightcome at the end of the test battery, order of testpresentation was random. Test data were entered into acomputer file by the psychometrician using an interactiveprogram. There were several double—checks to the data,including that out of bounds values were specified in thecomputer program. The 439-item NMPI was administered toeach subject as part of the test battery; the onlydifference from standard NNPI administration was that55subjects filled out computer cards. These cards were thencopied to disk by means of an optic scanner.Development of ScalesSigns and Symptoms (Sx)a) Content DomainFor this scale, two raters chose items from the MNPIwhich were judged to reflect true physical symptoms of MS orrelated health concerns. Raters were the Ph.D. candidate,Eleanor Boyle, who based selections on readings about MSsymptomatology, and Dr. Campbell M. Clark, a psychologistexperienced in neuropsychological testing. Only items onwhich raters agreed were included for subsequent analysis.In addition, items were deleted if missing responsestotalled >5%, or if all respondents answered similarly.Statistical analyses were performed using the StatisticalPackage for the Social Sciences, SPSS (Nie et a!. 1983),with the exception of the cluster analysis which was doneusing UBC:CGROUP (Lai, 1982).b) Item Analysis, Reliability and ValidityAn item analysis was done and any item was deletedwhich correlated with the overall scale with a coefficientr<O.15. After deletion of any such items, the scale wasanalysed for internal consistency, using Cronbach’s alpha56(Cronbach 1951). To ensure that this estimate was maximal,the analysis was done over the entire sample, MS and normalsubjects (Guttman 1945). As stated, items were initiallyselected on the basis of agreement between two raters.Subsequently, an analysis was done to determine whether thescales indeed distinguished the MS from the controlpopulations. To answer that question, the two populations’scores on the Sx scale were compared by means of a t—testfor difference in means, a Hartley’s F—maximum ratio fordifference in variance, and a Kolmogorov—Smirnov Z—test(Hays 1988) for a comparison of the two distributions.Classification rates were also determined by means of adiscriminant analysis.Distress (Ds)a) Content DomainMMPI items were chosen in the same manner as for the Sxscale, but to reflect patients’ psychological and emotionalhealth separate from MS-related physical symptoms.b) Item Analysis, Reliability and ValidityItem analysis, reliability and validity tests were doneas for the Sx scale.57Kurtzke (K)The Expanded Disability Status Scale (EDSS) was used asan objective measure of disability. Though in realityrecognized as somewhat subjective, and as a measure ofimpairment (clinical signs) more than disability(limitations to activities) (Willoughby and Paty 1988), theEDSS is the best-known assessment system for MS (Matthews etal. 1985). The EDSS rates patients’ overall physicalstatus, based on scores in Functional Systems (FS):pyramidal, cerebellar, brain stem, sensory, bowel andbladder, visual, cerebral/mental, and other/miscellaneous.Correlations were calculated between each pair ofmeasures: K and Sx, Sx and Ds, and K and Ds.Cluster AnalysisA z—score transformation was used to standardize thethree scales to the same underlying metric. Hence, eachscale had a mean of 0.0 and a standard deviation (SD) of1.0. This transformation ensured that each scale wasequally weighted for variance in the cluster analysis. Thismathematical transformation did not change thecharacteristics of the scales (e.g., ratio for Sx and Ds,ordinal for K) nor the shapes of the distributions, butsimply put those scales on a different metric (Bernstein1988). A cluster analysis was done using Ward’s method58(Ward 1963). Cluster analysis was chosen as a desirabletechnique because it can take several variablessimultaneously into account when judging individuals as moreor less “alike.” Cluster analysis is also designed forordinal scales (Jam and Dubes 1988) such as the Kurtzke,and for non—normal distributions of data (Morris et al.1981). Ward’s method was employed in this study because itis of the hierarchical agglomerative type most frequentlyused in scientific research (Morris et al. 1981), andbecause it has been shown to yield more accurate clusters,in general, than three other cluster analysis techniques:single linkage, complete linkage, and average linkage(Blashfield 1976), or at least to be among superiortechniques displaying high values of coefficient kappa,which describes improvement in subject classificationthrough use of the cluster analysis rather than randomassignment (Golden and Meehl 1980). Ward’s method is not assensitive to outliers as are other cluster analysis methods(Milligan 1980). This method is also biased in favor ofidentifying spherical clusters (Cormack 1971; Blashfield1976), which in this study refers to the Severity groups. Itshould be noted that the stability of a cluster solutiondepends on the reliability of the tests on which subjectsare being clustered.In the cluster analysis, the procedure was as follows:scores were compared among the 99 MS subjects on the threescales/variables (Sx, Ds and K). Distances were computed59between each variable for every possible two—subjectcombination. Subjects with the smallest squared—sumdistance between points (therefore the most similarity onthe three variables) were clustered into a group. So, forexample, if two subjects had scored equally on all threemeasures, they would have had zero distance between points,that squared sum would have been zero, and the two wouldhave been collapsed into one cluster. By this sequentialprocess 99 clusters became 98, then 98 became 97, until allsubjects were clustered into one. Selection of the optimalcluster solution (number of clusters) was based onstatistical criteria, namely the minimization of within-group variability and maximization of between—groupvariability, with the additional goal of as few clusters aspossible.To determine the homogeneity of the clusters, adiscriminant analysis was done with scale scores aspredictors (dependent variables) and cluster membership ascriteria (independent variables). The discriminant analysisshowed the efficacy of the cluster solution in its overallcorrect classification rate of subjects, since thatstatistic provided an estimate of the degree of overlap ornon—overlap among clusters.The z—scores were graphed, and graphical comparisonswere made between empirically—derived clusters, and thehypothesized ideal groupings discussed in Chapter One(figure I). It should be noted, however, that the analysis60operated without reference to any hypothesized profiles, andthat scores on the three measures are continuous, so thatany of a large number of “shapes” could have resulted fromthe data.Similarly—shaped clusters were then collapsed intoprofile groups. Severity—related clusters were defined asthose for which the range of scores on Sx, K and Ds scaleswas less than or equal to 1.0, the SD of each scale.A discriminant analysis was repeated on the newgroupings (of clusters into profiles) to ensure that nosignificant amount of information was lost due to thegrouping procedure. Because clusters making up anyparticular profile differed in severity, it was necessary tocorrect mathematically for severity before performing thediscriminant analysis. Correction for elevation or severityis a common procedure in profile analysis (Bernstein 1988).Profile groups were then compared on numerousdemographic and disease—related variables.Relations Between Groups and Measures of DiseaseMagnetic Resonance Imaging (MRI)MRI scans were rated by a radiologist for location,size and shape of lesions. The radiologist was not informedof the disease status of individuals before rating theirscans. However, given the nature of MS lesions, it cannot61be said that the radiologist was truly blind to whether asubject was MS or control. Presence or absence of lesionswas recorded on a standard form specifying possible lesionsites, with provision for the recording of otherabnormalities and related observations. These data werethen entered into a line file using an interactive programwritten for this purpose.a) Reduction of the Number of Lesion Sites to be ExaminedCriteria for inclusion of possible lesion sites was asfollows. The initial list of sites was, by design,overinclusive and consisted of 50 anatomical regions. Toreduce that number, the following regions were not analysed:(1) those at the midbrain or below, because they are lesslikely than higher regions to mediate cognitive andaffective behaviours; (2) those with a relatively highincidence of lesions in this sample of MS patients (>65%),because such regions were not likely to shed light oncluster—specific problems, and (3) those with a relativelylow incidence in this sample of MS patients (<5%).The aim of these criteria was to reduce the number oflesion sites as much as possible without loss of criticalinformation, thereby reducing the number of chancecorrelations between lesion sites and clusters. Reducingthe number of sites decreased the experiment—wise errorrate, namely the probability of encountering a Type I errordue to the number of tests of significance performed.62b) Relation of Psychological Profiles to Extent of LesionsFor each profile, an addition was made of the totalnumber of sites with lesions. This measure was distinctfrom total lesion burden or lesion load, since it did notattempt to take into account lesion size. It could insteadbe called extent of lesions, or extent of biologicaldisease. Statistical analysis entailed the following. Aone—way analysis of variance was done across profile groups,to answer the question: do any profiles differ from othersin number of sites with lesions? If a significant F-valuewas found, the groups were compared using Tukey’s ,posteriori pair-wise comparison technique (Hays 1988).c) Relation of Psychological Profiles to Lesion LocationThis analysis aimed to identify any anatomically-specific lesions which might contribute to a psychologicalprofile. Chi—squared analyses were done, site by profile,to answer the question: for a given site, do a large numberof members of any one profile have lesions there, whilemembers of other profiles do not? Ideal data were firstconceptualized, as illustrated in figure II, with groupingsbased on results of the cluster analysis. The figure showsa site in which all Depressed patients, but no others, havea lesion. Such a situation was both intuitively andstatistically significant; chi—squared analysis showed anassociated p-value 0.0001. It illustrates the simple63FIGURE IIMRI Lesion-location analysis:hypothesized ideal dataAn example of ideal data, in which all members of one profile(14/14 Depressed patients) have lesions in the site inquestion, and in which zero members of other profiles havelesions in that site. p < 0.0000‘Depression’ Denial ExagSom. SeverityPresence 14 0 0 0oflesionAbsenceof 0 32 22 31lesion64model in which any given lesion is associated with one, andonly one, profile.However, there was expected to be noise in the system,both “positive” (>0% of members of other profiles showinglesions in the site in question) and “negative” (<100% ofmembers of the profile in question showing lesions in thatsite). Sources of such noise could have included thefollowing:Positive. (1) MRI lesions can be clinically silent.This is particularly true for the transient lesions whichhave been revealed by serial scanning, and which arefrequently asyxnptomatic (Isaac et al. 1988); (2) lesionsjudged to be in the site in question may in fact be locatedslightly differently. Due to the size of sites, any twolesions both deemed to be in a particular anatomical site intwo different patients may not always be in precisely thesame nucleus or fiber tract.Negative. (1) Some members of a profile may bereactive, not organic; (2) not all MS patients, even thosewith CDMS, show detectable MRI abnormalities (Paty et al.1988)The question therefore arose: how close must data cometo the ideal, to be considered strong evidence for anorganic basis to the profile? The data in Appendix Baddress the question: how many members of a profile musthave a given lesion, when there is no noise from otherprofiles, for statistical significance? These data65illustrate that when even a small number of members of oneprofile, but not other profiles, show lesions in a givensite, a significant difference can be found.However, anticipating the existence of “positive” noisefrom other profiles, further data in Appendix B address thequestion: with varying degrees of noise how many members ofone profile must have a given lesion for statisticalsignificance?Based on the chi-squared calculations in Appendix B, itwas suggested that an alpha O.005 should constitutestatistical significance for the MRI lesion—locationanalysis. This alpha level avoids the potentially highfalse-positive (Type I error) rate which would be associatedwith a higher alpha.In the analysis of groups by lesion location, it wasdecided that outcomes would be examined which (1) showed astatistically significant difference among the groups for asite, and which also (2) were conceptually interpretablewithin the hypothesis that lesions can contribute to apsychological profile. Conceptual interpretability requiredthat (1) members of one and only one psychological profilestood out from others; (2) a profile differed from others bythe presence, not by the absence of a lesion. Presumably,it is the existence rather than the non—existence of alesion which can contribute to a psychological abnormality.Therefore, when chi—squared tests revealed statisticallysignificant differences between profiles, only those with66the indicated directionality were to be considered, and onlythose in which one profile was distinct from others.As mentioned, it was considered possible that somemembers of any profile had a reactive rather than an organicbasis to their behaviour. Intuitively, this idea issensible: in a group of severely depressed individuals, somemay possess that profile because of lesions, while othersmay possess it because their apparently mild disability hasa disproportionate impact on their lives — either real orperceived— as in the case of a pianist with minor sensoryloss to the hands. Therefore, as another step in theanalysis, for any one profile, patients with commoncharacteristic lesions were to be separated from thosewithout such lesions. Members of one or the other subgroup(with lesions or without) may have displayed commoncognitive problems which members of the other subgroup didnot.Cognitive measuresa) Selection of Tests to be AnalysedTests were chosen which examine a wide variety ofcognitive functions, and which have been shown to detectimpairment in MS (Peyser et al. 1980a; Heaton et al. 1985;van den Burg et al. 1987; Minden et al. 1990; Kionoff et al.1991). As well, tests were chosen which have minimal motorand sensory components (Tables 1 and 2), and which therefore67isolate cognitive functions as much as possible, so that anyof the patients’ motor and sensory impairments would notconfound results.b) Relation of Psychological Profiles to CognitivePerformanceThe profiles were divided into subgroups based on MRIresults, separating those which appeared to have anorganically-based profile from those which appeared to havea reactively-based one. Figure III illustrates that thisanalysis could potentially be done with eight groups,depending on the outcome of the NRI analysis. Unlike in thecase of the MRI analysis, data from normal controls wereexamined, so that controls formed one group. In the MRIanalysis, data from controls was not examined due to the lownumber of lesions in these individuals. However, in thecase of cognitive data there was variability amongindividuals. Therefore, it was useful and important tocompare MS patients’ scores with those of normals.As in the case of MRI lesion data, profiles wereexamined test—by—test to see whether scores on specifictests were related to profile membership. Ultimately, forevery cognitive test, the question was asked: do members ofone profile produce scores significantly worse than membersof other profiles? Ideal data were conceptualized and areillustrated schematically in figure IV, suggesting68FIGURE IIIPotential profile groups for neuropsychological analysisIf MRI results indicate that any of the original four profilegroups have organic and reactive subgroups, then analysis ofcognitive data will be done with up to eight subgroups, asshown.1. ‘Depressed’, organic2. ‘Depressed’, reactive3. Denial, organic4. Denial, reactive5. Exaggerated Somatic, organic6. Exaggerated Somatic, reactive7. Severity8. Normal controls69FIGURE IVNeuropsychological data analysis: hypothesized ideal dataAn example of ideal data, in which all members of one profilehave low performance on a cognitive test, while members ofother profiles have either medium or high performance on thattest. Underlining indicates that those groups are notsignificantly different.(low performance on cognitive test) < (high performance)ANY ONE PROFILE GROUP ALL OTHER MS GROUPS CONTROLSORANY ONE PROFILE GROUP ALL OTHER MS GROUPS CONTROLS70situations in which one profile group stands out from otherswith significantly lower performance on a cognitive test.Before test—by—test analysis, however, it was firstnecessary to do an omnibus test for overall significance, inthe form of a Multivariate Analysis of Variance (MANOVA)(for <8 profile groups as independent variables, and 19continuous dependent variables in the form of scores on theselected cognitive tests). If the MANOVA resulted in asignificant F-value, ANOVA5 could subsequently be performedon individual cognitive tests. If Analysis of Variance(ANOVA) resulted in significant F-values (p 0.05) forcertain cognitive tests, then Tukey’s a posteriori testscould be performed to see where between—group differencesexisted. Of available tests, Tukey’s was suitable becauseit maintains alpha over all pair—wise comparisons (Winer1962), thereby controlling for experiment—wise error.The number of subjects in this study was large byclinical research standards. Nevertheless, theneuropsychological variables in the study were alsosufficiently numerous that the subject—to—variable ratio wassmaller than multivariate analysts would advise. However,the validation study served as a check in this regard.71CHAPTER FOURRESULTSSubject CharacteristicsThe subjects consisted of 99 patients and 56 normalcontrols. Of the 99 patients, 69 (69%) were female and 30(30%) were male. The patients had a mean age (standarddeviation) of 36.3 (7.9) years, and a mean education of 13.8(2.3) years. Mean age of first symptoms for these patientswas 25.5 (7.1) years, while mean age of diagnosis was 30.6(8.4). Number of relapses since the diagnosis averaged 5.3(2.6), so that the average number of relapses per year forthis group was 0.93. Results of neurological examinationsfurther suggest that this is an MS group with relatively mildphysical problems. Mean (SD) Kurtzke EDSS for all patientswas 2.03 (1.38) (on a scale of 0.0—10.0 in which 10.0represents death from MS). No patients scored higher than6.0, which was part of the selection criteria. A summary ofpatient characteristics is given in Table 3.Controls were well-matched with MS patients ondemographic variables, as shown in Table 4. T—tests for thecontinuous variables, and chi—squared tests for thecategorical variables, showed that there were no significantdifferences between the two groups on all but one of thesevariables. The only variable on which patients differed72TABLE 3Characteristics of MS subjectsNumber of subjects = 99Sex (female/male) = 69/30SD7. meanage 36.3education (yrs.) 13.8age of first symptoms 25.5age of Dx 30.6number of relapses since Dx 5.3number of relapses/year 0.93EDSS 2.0373TABLE 4Comparison of MS subjects with normal controlson demographic variablesvariable MS Controlssex (% female) 69.0 66.7age 36.3 35.9education (yrs.) 13.8 13.8marital status:married 51.5 50.9single 19.2 30.2separated/divorced 28.3 18.9occupation:prof/semi-prof 22.0 22.2managerial 7.0 9.3clerk/skilled 49.0 46.3unskilled 14.0 14.8homemaker 1.0no occupation 3.0student 4.0 7.4employment status:full—time 36.4 57.1part-timeby choice 24.2 21.4unemployeddue to health 19.2 -other 20.2 21.574significantly from controls was current employment status(chi—sq=18.54; df=9; p 0.029).Derived Measures: Sx and DsThe MNPI items used on the Sx and Ds scales are given inAppendix C. Final numbers of items were 31 for the Sx scaleand 47 for the Ds scale. It was not necessary to delete anyitems for which more than 5% of responses were missing, sinceno such items were found. However, one potential Sx item hadto be deleted for lack of variability. All subjects, both MSand controls, responded positively to “I feel tired a gooddeal of the time”, which may have said more about modern lifethan about MS. The item analysis showed that the vastmajority of items on each scale correlated highly with theoverall scale; however, three items on each scale were deletedbecause the correlation with the total score was less than0.15. Following the item analysis, internal consistencycoefficients were over 0.9 for both scales (Sx: r=0.901; Ds:r=0.900)Tests for criterion-related validity showed that the Sxscale separated the two samples well. On the Sx scale, MSpatients scored significantly higher than controls (MS:mean=l3 .9, SD=5 .9; controls: mean=4. 3, SD=3 .5, t=l2 .83;p < 0.001). The Kolmogorov—Smirnov z-test showed that scoreson Sx were normally distributed for MS but were significantlydifferent from normal for controls (z=1.6; p < 0.01). The75discriminant analysis also revealed a correct classificationrate for the Sx scale of 83.4% of all subjects.The Ds scale also separated MS subjects from controls,with MS patients scoring significantly lower (indicating moredistress) than controls (MS: mean=35.4, SD=8.l; controls:mean=38.9, SD=6.7; t=—2.72; p < 0.005). Scores on Ds werealso normally distributed for MS patients but weresignificantly different from normal for controls (z=l.5,p 0.03). The discriminant analysis showed a classificationrate of 58.9% on Ds. These data are summarized in Table 5.Correlations between pairs of measures, calculated for MSpatients (n=99) were as follows: K and Sx: r=0.30, p 0.002;K and Ds: r=0.lO, p < 0.15; Sx and Ds: r=0.47, p 0.001.Correlations were also calculated between Sx and FunctionalSystem (FS) scales, and between Ds and FS scales, and nonewere significant.Cluster AnalysisA 10-cluster solution was statistically optimal. Forthat solution, 81.5% of total multivariate variability wasbetween groups, with the remaining 18.5% within groups. Thediscriminant analysis of this solution yielded a 97.98%correct classification rate, suggesting clear separationbetween the clusters in multivariate space. None of theclusters contained fewer than five subjects.76TABLE 5Sx— and Ds—scale analysesSx Dsreliability coefficient 0.9 0.9Mean score:MS patients: 13.9 35.4Controls: 4.3 38.9t-value for difference: 12.83 —2.72associated p—value: 0.00l O.005Classification rate of scale: 83.4% 58.9%77Empirically—derived clusters are graphed, as can be seenin figure V, which also shows the number of subjects in eachcluster. The 10 clusters were collapsed into four profilegroups based on visual similarity. After a correction forseverity, a discriminant analysis on the four—profile groupingrevealed that the classification rate remained 97.98%,indicating that no information was lost in the grouping of 10clusters into four profiles.1. The ‘Depression’ profile contained 14 patients, intwo clusters. As can be seen from the graphs (figure V),‘Depressed’ patients scored high on Ds reflecting a distressout of proportion to their relatively low scores on objectiveassessment of disability (K). These patients are alsodistressed despite their acknowledgement, as evidenced bytheir low Sx scores, that they do not have many physicalproblems.2. The Denial profile contained 32 patients, in threeclusters. These patients were characterized by low scores onboth of the self-rating scales (Ds and Sx) but relatively highscores on objective assessment of disability (K), indicatingthat they do have physical problems, but report that they donot have problems either physical or psychological.3. The Exaggerated Somatic profile contained 22patients, in two clusters. This group was characterized byrelatively high scores on Sx, compared to their scores on K,signifying that these patients report their physical problems78FIGUREVResultsofclusteranalysis:psychologicalprofilesinMS>•1-DU)-o0)CCl)ci)I()‘Depression’(n=1 4)•Cluster#8;n=51•Cluster#4;n=16oCluster#1;n=6I-22•Cluster#9;n=8oCluster#2;n=16•Cluster#7;n=7Denial(n=32)20•Cluster#3;n=7oCluster#1.0;n=5•Cluster#6;n=20210—1-2210—1-2SxKDsSxKDsIIIExag.Somatic(n=22)>4—-o(C,)-oC)CU)0CSeverity(n=31)oIN1-2SxKDsSxKDsas considerably worse than do their neurologists. Scores onDs were slightly elevated for some patients.4. The Severity profile contained 31 patients, in threeclusters. Characteristic of this group was that the meanz—score for a group on all three measures was within onestandard deviation of one another. This score indicates thatthese patients’ assessments of their physical problems weregenerally concordant with assessments by their neurologists,and that patients with few problems experienced a small amountof distress while patients with many problems experienced moredistress.Analyses of variance on profile groups by educationshowed no difference among groups. ANOVA on groups by ageshowed some differences (F395=2.72; p 0.049). The mean agefor ‘Depressed’ patients was lowest (31.5 years), forSeverity-related patients was 35.4 years, for ExaggeratedSomatic patients was 37.4 years, and for Denial patients was38.1 years. A Tukey’s a posteriori test showed a significantdifference between ‘Depressed’ and Denial patients but notbetween other pairs of groups.Several related analyses were also done. ANOVA5 showedthat groups did not differ on any of (1) age of onset of firstsymptoms, (2) age of diagnosis, (3) number of years sincediagnosis. Groups did differ on number of years since onsetof first symptoms, with Deniers (who are also oldest) havinglived with their symptoms slightly longer than other groups.These data, for mean (SD) number of years since onset of first80symptoms, are as follows: ‘Depressed’: 6.6 (6.2); ExaggeratedSomatic: 11.9 (7.1); Denial: 13.5 (7.3); Severity: 8.8 (5.6).Analyses of profile groups by sex showed that 13/14 (93%)of ‘Depressed’ patients were female, although the sample was69% female (69/99). Chi-squared analysis showed that thisoverrepresentation was not significant (p 0.056). Womenwere not overrepresented in any other group, includingExaggerated Somatic.Scores on the Beck Depression Inventory were markedlydifferent among groups, as shown by an ANOVA (p 0.001) with‘Depressed’ patients scoring higher than members of otherprofiles, thus supporting criterion—related validity for the‘Depressed’ profile.MRI AnalysisRelation of Psychological Profiles to Number of Sites WithLesionsAn analysis of variance was done to compare all fourprofile groups on total number of sites with lesions, of the50 anatomical sites. The groups were not significantlydifferent (F393=2.14; p < 0.101). Means (SDs), for number ofsites with lesions, were as follows: ‘Depressed’: 16.3 (8.2);Denial: 13.6 (5.8); Exaggerated Somatic: 11.5 (6.4); Severity:11.7 (6.2), as presented in Table 6.81TABLE 6Profiles did not differ significantlyon number of sites with lesionsF393 = 2.14 p < 0.101group mean SD‘Depression’ 16.3 8.2Denial 13.6 5.8Exag. Somatic 11.5 6.4Severity 11.7 6.282Relation of Psychological Profile to Lesion LocationFor the initial analysis, the list of 50 sites wasreduced based on criteria discussed in Chapter Three. A totalof 23 sites were excluded: nine because they were at or belowthe level of the midbrain, seven because more than 65% of MSpatients had lesions there, and another seven because fewerthan 5% of MS patients had lesions there. Therefore, 27 siteswere analysed. (Data on the percentages of patients andcontrols who had lesions in given locations are in Table 7.)Site-by-site chi-squared analysis showed that 26 of the 27sites did not satisfy the statistical criterion of p 0.005.One site, the gray matter/white matter junction of theright parietal lobe, did satisfy the criterion. In that site,half the ‘Depressed’ group (7/14) had lesion(s), while smallerpercentages of other groups had lesion(s) there. These dataare summarized in Table 8. The largest percentage of anyother group with a lesion there was 5/24 (20.8%) for Severitypatients.These data prompted several further analyses. One set ofanalyses compared the two ‘Depressed’ subgroups, one whosemembers had the lesion and one whose members did not, ondemographic variables (age, sex, and education) as well asseverity of impairment/disability (EDSS) and total number ofsites with lesions. Another set of analyses compared the twogroups of patients who had lesions in that site, one group‘Depressed’ and one group composed of all non—’Depressed’83TABLE 7Percentage of subjects with lesions in each brain area.Asterisks indicate sites used for chi—squared analysis.Slice Region MS ControlsSupraventricular: 1) Frontal: R 17.5* 02) L 17.5* 03) Frontal/Parietal: R 18.6* 04) L 19.6* 05) Parietal: R 30.9* 06) L 3Q9* 0Periventricular: 7) Frontal Horn: R 82.5 21.48) L 83.5 26.89) Occipital Horn: R 77.3 17.910) L 77.3 16.111) Temporal Horn: R 40.2* 1.812) L 43•3* 013) Parietal Body: R 84.5 5.414) L 83.5 3.6Deep White: 15) Frontal Horn: R 17.5* 1.816) L 12.4* 017) Occipital Horn: R 6.2* 018) L 4.1 019) Temporal Horn: R 3.1 020) L 5.2* 021) Parietal Horn: R 39.2* 022) L 42.3* 1.8Internal Capsule: 23) R 16.5* 024) L 15.5* 0Gray/White 25) Frontal: R 22.7* 0Junctions: 26) L 17.5* 027) Parietal: R 17.5* 028) L 19.6* 029) Occipital: R 5.2* 030) L 2.1 031) Temporal: R 11.3* 032) L 11.3* 0Deep Gray: 33) Insula: R 9•3* 034) L 6.2* 035) BG: R 2.1 036) L 1.0 037) Thalamus: R 1.0 038) L 2.1 0Brain Stem: 39) Midline: 25.8 1.840) Cerebellum: R 20.6 1.841) L 10.3 042) Pons: R 23.7 043) L 20.6 044) Mid Brain: R 14.4 045) L 16.5 046) Medulla: R 10.3 047) L 7.2 0Corpus Callosum: 48) Body 71.1 14.349) Genu 35.1* 050) Splenium 39.2* 1.884TABLE 8Gray matter-white matter (GM-WI.!) junction of the rightparietal lobe: chi—squared analysis(chi—sq: 13.34; p . 0.004)‘Depression’ Denial Exag. Somatic Severitypresence 7absence 74 1 528 21 2485patients with that lesion, on Beck Depression Inventoryscores, on cognitive impairment as reflected in scores onneuropsychological tests, and on finer detail of anatomicallocation in the right parietal lesion site. Results of theseanalyses are summarized in the section entitled Post HocAnalyses.In summary, analyses of MRI lesion location data showedthat the majority of patients in three out of four profilegroups were not distinguished by having lesion(s) inparticular sites. A subset of the ‘Depressed’ group wasdistinguished by lesion(s) in the right parietal lobe.Neuropsychological AnalysisNeuropsychological analysis was done on six groupssubdivided according to the MRI data.(1) ‘Depressed’—lesion(2) ‘Depressed’—no lesion(3) Denial(4) Exaggerated Somatic(5) Severity(6) Normal controlsA MANOVA produced a significant multivariate F (F=l.68df=(120,570); p < 0.001), indicating overall differencesbetween the six groups on 22 measures (19 neuropsychologicaltests and three IQ measures)86Test-by-test ANOVAs showed: (1) no significantdifferences at p 0.05 on 12 neuropsychological measures, oron Verbal IQ; (2) significant differences at p 0.05 onPerformance IQ (p < 0.0001) and Full—scale IQ (p 0.0022);(3) significant differences on seven neuropsychological tests,as shown in Table 9.An examination of the data showed that the ‘Depressed’group with the lesion performed worse than other groups on allseven of the above tests for which significant differenceswere found. Tukey’s comparisons were done, and in some casesshowed significant differences between the ‘Depressed’ groupwith the lesion and other groups, as illustrated in figure VI.Differences were in some cases not significant because of thesmall size of the two ‘Depressed’ subgroups (each n=7).Group scores on the above—mentioned sevenneuropsychological tests are illustrated graphically in figureVII. For this graph, groups’ scores were standardized using az—score transformation, so that each test had a mean of 0.0and an SD of 1.0. As can be seen in figure VII, scores forthe ‘Depressed’ group with the lesion were consistently low.As can also be seen, normal controls’ scores are generallyrelatively high, though not always higher than all MS groups.The ‘Depressed’ group without the lesion had variable scores -sometimes low, but in several cases higher than those ofnormals. The ‘Depressed’ group without the lesion performedbetter than normal controls on five of the 19 tests: Paired87TABLE 9Neuropsychological tests on which ANOVA5 showed significantdifferences among groupsTestPaired Associate LearningObject Assembly (WAIS)Word FluencyB. Visual RetentionBlock Design0.00510.01000.01950.0340F—value p-valueF5 150=6.33 0.0000F5150=4.24 0.0013F5150=3.74 0.0032F5 150 . 50F5 15cr3 . 14Memory for Objects F5,150=2.79Speech Perception F5150=2.4988FIGURE VIResults of Tukey’s tests for neuropsychological tests on whichANOVA5 showed significant differences among groupsPairs of groups which were significantly differenton Tukey’s (p < 0.05) are denoted by (*)‘Depressed’ with right parietal lesion‘Depressed’ without right parietal lesionExaggerated somaticDenialSeverityNormal controlsDL DEN EXAG 5EVDL =DNL =EXAG =DEN =SEV =Pairs:OA:N DNLDL * * *DEN *EXAGSEV *N * *DNL *DL DEN EXAG SEV N DNLDL *DENEXAGSEVN *DNL89FIGURE VI continued:Results of Tukey’s tests. Groups which were significantlydifferent (p < 0.05) are denoted by (*)‘Depressed’ with right parietal lesion‘Depressed’ without right parietal lesionExaggerated somaticDenialSeverityNormal controlsDL DEN EXAG SEV N DNLDEN *E XAGSEVN *DNLBVRT:DL DEN EXAG SEV N DNLDLDENEXAG *SEVN *DNLDL =DNL =EXAG =DEN =SEV =N=WF:DL90FIGURE VI continued:Results of Tulcey’s tests. Groups which were significantlydifferent (p < 0.05) are denoted by (*)DL = ‘Depressed’ with right parietal lesionDNL = ‘Depressed’ without right parietal lesionEXAG = Exaggerated somaticDEN = DenialSEV = SeverityN = Normal controlsBD:DL DFN FXAG SFV N DNLDL * *DENEXAGSEV *N *DNLSpPer: No significant differences on pair—wisecomparisonsMemOb: No significant differences on pair—wisecomparisons91FIGURE VIIComparison of groups on neuropsychological testsThe graph shows that scores were consistently low for the ‘Depressed’ group with thelesion. The tests illustrated are those on which ANOVAs revealed significantdifferences between groups: Paired Associate Learning (Pairs), Object Assembly(OA), Word Fluency (WE), Benton Visual Retention Test (BVRT), Block Design(BD), Speech Perception (SpPer), and Memory for objects (MemOb).10—1-20 Pairs CA WF BVRT0 Depressed’- no lesion0 Normals• Severityo Exag Somatic• Denial• ‘Depressed’ - lesionBD SpPer MemOb92Associate Learning, Categories, Speech Perception, PictureCompletion, and Memory for Objects.The ‘Depressed’ group with the lesion performed worst ofall groups on all three IQ measures and on the majority (18out of 19) of other neuropsychological tests as well. Scoresfor the ‘Depressed’ group with the lesion showed moreimpairment than other MS groups on all tests except DigitSpan, on which the other ‘Depressed’ group without the lesionperformed worst.An examination of group means on all neuropsychologicaltests showed that MS patients generally performed worse thannormals.For descriptive purposes, correlations betweenneuropsychological tests were calculated, and are shown inTable 10. Correlations were calculated for those seven testson which ANOVAs showed significant differences among thegroups.Post Hoc AnalysesBecause the MRI lesion-location analysis distinguishedseveral groups of patients on the right parietal lesion site,it was considered appropriate to do further analyses forcharacterization of these groups. Three groups weredistinguished which warranted further examination: (1)‘Depressed’ patients with the right parietal lesion (n=7);93TABLE 10Correlations between tests on which ANOVAs showedsignificant differences among groupsMS/ControlsPairs 1.00 .38/ .27/ .43/ .33/ .37/ .47/.25 .21 .29 .34 .29 .21CA 1.00 .27/ .48/ .61/ .23/ .23/.13 .21 .55 .10 .03WF 1.00 .18/ .36/ .16/ .23/.07 .13 .05 .31BVRT 1.00 .66/ .30/ .29/.48 .21 .11BD 1.00 .31/ .18/.20 .17SpPer 1.00 .12/.23MemOb 1.00Pairs CA WF BVRT BD SoPer MemOb94(2) ‘Depressed’ patients without that lesion (n7); (3) Non-‘Depressed’ patients with that lesion (n=l0).Comparisons Between the Two ‘Depressed’ GroupsThe two ‘Depressed’ groups, one with the lesion and theother without the lesion, were compared on demographicvariables. They were compared on age and education (t—tests),and on sex (chi—squared). No significant differences werefound.The two groups were compared on severity of clinicalsigns (EDSS score), and no significant difference was found.Mean (SD) scores were: ‘Depressed’ with the lesion: 1.07(1.2); Non-’Depressed’ with the lesion: 1.00 (0.58).The two groups were also compared on total numbers ofsites with lesions. Means were as follows: ‘Depressed’ withthe lesion: 22.3 (6.1), ‘Depressed’ without the lesion: 10.3(5.1), as shown on an ANOVA across these groups as well asacross two further groups: Non—’Depressed’ with the lesion:20.1 (5.0), and all other MS patients: 11.3 (5.4). As can beseen, it was not the two ‘Depressed’ groups who were similaron this variable. Rather, it was the two groups with thecharacteristic lesion, both of whom had a large number ofother lesions as well. The ANOVA showed a significantdifference among the groups (F393=15.59; P 0.0001); aTukey’s test showed that the ‘Depressed’ group with the lesionand the Non—Depressed group with the lesion were both95significantly different from the two other groups. Data onthis variable are summarized in Table 11.These data therefore show that ‘Depressed’ patients withthe lesion have considerably more lesions overall than do‘Depressed’ patients without that lesion. Other patients withthat lesion also have a large number of lesions overall.Comparisons Between the two Groups With the Lesion,‘Depressed’ and Non—’Depressed’These two groups were compared on the Beck DepressionInventory (BDI), to assess whether both groups were depressedin a way not captured by the original profile definitions.The two groups were compared with all other MS patients, andwith normal controls. All MS groups were found to have someBDI—measured depression compared with normals. Mean scoresare summarized in Table 12. (BDI scores measure amount ofdepression as follows: O—4=none or minimal; 4—7=mild; 8—15=moderate; >16=severe (Beck and Beamesderfer 1974)). ATukey’s test showed differences between normal controls andall MS groups except Non—’Depressed’ with the lesion. Becausethese groups contain small numbers of patients, significancetests may not have the power to show differences, in this casebetween normal controls and Non—’Depressed’ with the lesion.These data indicate that all MS patients are more depressedthan normal controls, and that patients who are Non—‘Depressed’ with the lesion have virtually no more depression96TABLE 11Comparison of groups on total number of sites with lesions,for groups subdivided in reference toright parietal lesion(F3,9315.59, p < 0.0001)mean SD‘Depressed’with lesion(n = 7)‘Depressed’without lesion(n = 7)Non—’Depressed’with lesion(n = 10)22.3 6.110. 3 5.120.1 5.0Other MS(n = 75)11. 3 5.497TABLE 12Comparison of two ‘Depressed’ Groups, Other MS Subjects, andControls, on Beck Depression Inventory (BDI)meanNormal Controls 3.66‘Depressed’ with lesion 11.4Non—’Depressed’ w/ lesion 7.1Other MS 6.598than most MS patients. The group that is Non-’Depressed’ withthe lesion also has less depression than does the group of‘Depressed’ patients with the lesion, although differencesbetween MS groups are not significant.As stated in the previous section, the ‘Depressed’ groupwith the lesion scored worse than other groups on mostneuropsychological tests (21/22). These data prompted thequestion: are all patients with the right parietal lesioncognitively impaired? ANOVAs were re-done on allneuropsychological tests on the following groups: ‘Depressed’with the lesion, Non-’Depressed’ with the lesion, other MS,and Normal controls. The Non—’Depressed’ group with thelesion was not noticeably impaired on any test. That is,scores for the Non—’Depressed’ with lesion group wereconsistently close to those of other MS subjects (except‘Depressed’ patients with the lesion whose neuropsychologicalscores were low) and no significant differences were foundbetween the Non—’Depressed’ with the lesion group and otherMS.It was also reasoned that the two groups with the lesion- one ‘Depressed’ and one not — might differ on finer detailof anatomical location of their lesions in the indicated rightparietal site. For example, the ‘Depressed’ group with thelesion may have lesions more anterior to, or more posteriorto, those of the Non—’Depressed’ patients. Therefore, avisual inspection was done of the original MRI scans, by thePh.D. candidate, together with the radiologist. There were no99obvious similarities within the ‘Depressed’ group, nor werethere any marked differences between the two groups onanatomical location within the site.100CHAPTER FIVEVALIDATION STUDY: METHODSSublectsFor the validation study, data were available from 43additional MS patients assessed after the initial study. Allpatients had clinically definite MS (CDMS) with a relapsing-remitting course, and satisfied the same criteria (detailed inChapter Three) as initial subjects. Data were also availablefrom 20 normal controls assessed after the 56 controls of theinitial study. All met the same criteria as did earliercontrols.Subjects were assessed in the manner described in ChapterThree, with a neurological examination, an interview, a seriesof MRI brain scans, and psychological tests.Demographic data were obtained, and comparisons were madebetween MS and controls, on age and education (t—tests), sex,marital status, occupation and current employment status (chisquared)Sx and Ds Scales: Reliability and ValiditySx and Ds scales, which had been analysed for reliability(internal consistency) on the initial 99 subjects, were re101analysed for internal consistency (Cronbach 1951) on thevalidation subj ects.Subsequently, Sx and Ds were re—assessed to show whetherthe scales distinguished the MS subjects from controls. Oneach scale, a t—test was done for a difference in meansbetween the two groups.Cluster AnalysisThe 43 MS patients were then cluster-analysed. First,raw scores were standardized to z—scores for the threemeasures (Sx, Ds and K) which formed the basis of the clusteranalysis. The analysis was then undertaken, using Ward’smethod (Ward 1963). Selection of the optimal cluster solution(number of clusters) was based on statistical criteria:minimization of within-group variability and maximization ofbetween-group variability, with the additional goal that thereshould be a small number of clusters and that each clustershould contain at least two members. Clusters with similarscores were collapsed into profile groups, and a discriminantanalysis done to ensure that the groups showed a highclassification rate. Group means were graphed.Profile groups were compared on age (ANOVA), sex, maritalstatus, occupation and employment (chi—squared).102MRI AnalysisIn light of results from the initial study, which showedthat groups differed significantly on number of lesions in theGM-WM junction of the right parietal lobe, the analysis wasrepeated on that site. Only that site was examined. It waspredicted that validation subjects with ‘Depressed’ profileswould show significantly higher numbers of lesions in thatsite than would subjects with other profiles.Chi—squared analysis was performed between profilegroups. For this analysis, the statistical criterion wasp < 0.05, rather than p < 0.005 as in the earlier analysis.There were several reasons for the use of a more liberal p—value in the validation study. Firstly, because only one chisquared analysis was done, experiment—wise error was not apotential problem. Secondly, group sizes were small in thevalidation study, so a higher alpha was needed for increasedpower.An added analysis checked the results of the initialstudy showing that groups do not differ significantly onnumber of sites with lesions. An ANOVA was done across groupson this variable.Neuropsychological AnalysisA further analysis was needed to validate theneuropsychological finding from the initial study, that is,103cognitive impairment among ‘Depressed’ patients who have theright parietal lesion.Cognitive tests were examined on which significant groupdifferences had been found in the initial study. Therefore,nine measures were analysed— two IQ (Performance IQ and Full-scale IQ) and seven neuropsychological measures: PairedAssociate Learning (Pairs), Object Assembly (OA), Word Fluency(WF), Benton Visual Retention (BVRT), Block Design (BD),Memory for Objects (MemOb), and Speech Perception (SpPerc).On all seven of these tests, in the earlier analysis the‘Depressed’ group with the lesion was most impaired; it wastherefore predicted that ‘Depressed’ members of the validationgroup (and particularly any ‘Depressed’ patients with theright parietal lesion) would have significantly greaterimpairment on these tests than would other patients.ANOVA5 were done on each of the nine measures. ForANOVAs which resulted in significant F-values (p 0.05),Tukey’s a posteriori tests were performed to see which pairsof groups were significantly different.104CHAPTER SIXVALIDATION STUDY: RESULTSSubj ectsOf the MS subjects, 33 (76.7%) were female and 10(23.3%) were male. The patients had a mean age (standarddeviation) of 34.2 (6.7) and a mean education level of 13.6(2.0) years. Mean age of onset of the disease for thesepatients was 26.3 (6.4). Results of the neurologicalexaminations showed that the mean (SD) Kurtzke EDSS scorefor this group was 1.87 (1.16). A summary of patientcharacteristics is given in Table 13. Characteristics ofthis patient group are similar to those of the MS patientgroup in the initial study.Comparison of MS subjects and normal controls (n20) ondemographic variables showed that the patients did notdiffer significantly from controls on age, sex, education,marital status, or occupation when employed. However, thegroups did differ on current employment status (chisquared12.6; df=6; p 0.049), with more MS than controlsunemployed for health reasons, as was the case in the mainstudy. On demographic variables (Table 14) the validationsample was very similar to the sample in the main study.105TABLE 13Validation study: Characteristics of MS subjectsNumber of subjects = 43Sex (female/male) = 33/ SD34.2education (yrs.) 13.6age of first symptoms 26.3EDSS 1.87106TABLE 14Validation study: Comparison of MS subjects and normalcontrols on demographic variablesvariable MSsex (% female) 76.7age 34.2education (yrs.) 13.6marital status:married 79.1single 16.3separated/divorced 4.7occupation:prof/semi-prof 20.9managerial 4.7clerk/skilled 58.1unskilled 9.3homemaker 2.3student/other 4.7employment status:full—time 37.2part-timeby choice 20.9unemployeddue to health 16.3other 25.6Controls85. 034.114 . 060.025.010. and Ds ScalesReliability analyses for internal consistency on Sx andDs scales resulted in correlation coefficients of 0.84 forSx, and 0.92 for Ds.T—tests showed that the scales separated the twosamples. On the Sx scale, MS patients scored significantlyhigher than controls (MS: mean=14.O, SD=5.8; controls:mean=4.l, SD=3.4, t=8.66; p 0.001). On the Ds scale, MSpatients’ scores were significantly different from those ofcontrols (MS: mean=32.8, SD=9.3; controls: mean=41.0,SD=6.2; t=-3.58; p < 0.001). Controls had higher Ds scoresin the raw data because that was the directionality assignedto the scale. In the case of Ds, the z—scores werereflected (i.e., x=—x) so that all three scales had the samedirectionality. These data are summarized in Table 15.Cluster AnalysisThe statistically optimal solution contained 10clusters. For that solution 86.5% of multivariatevariability was between groups, and the remaining 13.5% ofvariability was within groups. None of the clusterscontained fewer than two subjects.Profiles for empirically—derived clusters werecompared, and similar clusters collapsed into four profilegroups. A discriminant analysis on the four—profile108TABLE 15Validation study: Sx— and Ds—scale analysest—value for difference:p 0.001Sx Dsreliability coefficient 0.84 0.92Mean score:MS patients: 14.0 32.8Controls: 4.1 41.0associated p—value:7.12 —3.58p 0.001109grouping showed a classification rate of 95.35% (2/43subjects misclassified).Group means were graphed (figure VIII). As can be seenfrom these graphs, profiles were similar to those of theinitial study, and were consistent with the labels‘Depression’, Denial, Exaggerated Somatic and Severity.None of the four groups contained fewer than nine subjects.Cluster groups did not differ significantly on sex(chi-sq=4.77; df=3; p < 0.19), marital status (chi-sq=l9.23;df= 18; p < 0.38), occupation (chi—sq=20.61; df=21;p < 0.48), or employment status (chi—sq=21.91; df=18;p < 0.24).Analysis of profile groups by age yielded a significantF-value (F3,9=3.29; p 0.031). The mean (SD) age for‘Depressed’ patients was 29.2 (6.96), for Denial was 33.8(7.27), for Exaggerated Somatic was 36.3 (5.79), and forSeverity was 37.2 (4.29). A Tukey’s test showed asignificant difference at p 0.05 between ‘Depression’patients, and Severity patients. As in the main study, the‘Depressed’ group had a slightly lower average age than didother groups. However, the group with the highest averageage was not Denial, as in the main study, but Severity.MRI AnalysisAn analysis of variance compared the four profilegroups on number of sites with lesions, on all 50 anatomical110FIGUREVIIIValidationStudy:resultsofclusteranalysis2‘Depression’(n=10)2Denial(n=12)11.2O9-0.2590.10-1—1SxKDsSxKDsHHH2Exag.Somatic(n=12)2Severity(n=9)11-0.025--O.8SxKDxSxKDssites. The groups were not significantly different(F3,7=2.06; p 0.12). Means (SDs) were: ‘Depression’ 17.7(5.8), Denial 12.3 (7.1), Exaggerated Somatic 11.6 (5.1),and Severity 11.6 (8.4), as outlined in Table 16.Statistical analysis was done on the site (GM-WMjunction of the right parietal lobe) on which a significantdifference had earlier been found. A larger percentage ofmembers of the ‘Depressed’ group (4/10=40%) had lesions inthat site than did members of other groups (Denial=25%;Exaggerated Somatic=9%; Severity=0%). That difference wasnot significant by chi-squared analysis at p 0.05 (chisq=6.27; df=3; p < 0.099). However, there was clearly atrend. Kendall’s tau is also an appropriate statistic. Itrequires directionality of scale in one or more variables;in this case the dependent variable (whether or not anindividual had a lesion) can be directional. Thedistribution was significant using Kendall’s tau (tau=-0.277; p < 0.028). The data for groups at this site isshown as a matrix in Table 17.Neuropsychological AnalysisThis analysis was done on four groups: (1) ‘Depressed’with the lesion, (2) ‘Depressed’ without the lesion, (3) Allother MS subjects, (4) Normal controls. It was notnecessary to do this analysis on six groups, as in the mainstudy: the intent of this part of the validation study was112TABLE 16Validation study: Profiles did not differ significantlyon number of sites with lesionsF337 = 2.06 p < 0.123group mean SD‘Depression’ 17.7 5.8Exag. Soiriatic 11.6 5.1Denial 12.3 7.1Severity 11.6 8.4113TABLE 17Validation study: GM-WM junction of the right parietal lobe:chi-squared analysis(Kendall’s tau B = —0.277; p < 0.028)presence 4‘Depression’ Exag. Somatic Denial Severityabsence 61 211 809114to focus on the group of ‘Depressed’ patients with thelesion.One—way ANOVA5 on the two IQ measures and sevenneuropsychological tests showed the following results.Significant differences among groups were found onPerformance IQ (F359=4.9l; p 0.004); Full—scale IQ(F3,59=3.39; p 0.024) ; Object Assembly (F359=3..44;0.023); Benton Visual Retention (F3,59=2. 7; P < 0.039);Block Design (F359=3.03; p 0.036); and Paired AssociateLearning (F3,59=3.07; p < 0.035), but not on Word Fluency,Memory for Objects or Speech Perception. However, in noneof the Tukey’s comparisons were the scores of ‘Depressed’patients with the lesion significantly lower than others.Because the power of a statistical test increases withincreasing sample size, the small size of the ‘Depressed’group with the lesion (n=4) precluded findings ofsignificance in some tests. ‘Depressed’ patients with thelesion did worst of all groups on Full—scale IQ, BentonVisual Retention, Memory for Objects, and Paired AssociateLearning. Overall, however, ‘Depressed’ patients with thelesion did not stand out as being markedly more impairedthan other groups.115CHAPTER SEVENDISCUSSION AND CONCLUSIONSMethodological ConsiderationsBefore conclusions are drawn from this study, a numberof methodological issues should be reviewed. First, the MSsample in this study was deliberately restricted, so thatsubjects had mild clinical symptoms, a relapsing/remittingMS course, and no complicating factors such as otherdiseases or drug use. Because of this selective sampling,inferences can be drawn with confidence for MS patients withmild physical problems. However, results cannot beextrapolated to patients with more severe disease. Forexample, the finding that psychological response may belargely a reaction to stress for this sample does notpreclude the possibility of organically—based psychologicalresponse in the later stages of the disease. Further to theissue of sampling is the fact that patients for the studyall had EDSS 6.0. The chief reason for restricting therange in this manner was that patients with higher EDSSscores would more likely have had dementia, and motor orsensory problems, which could have invalidated their scoreson a variety of tests. As well, because relations between arestricted range of a variable, and other variables, can beattenuated, the fact that psychological profiles were116identified in this range is further evidence for theexistence of these profiles.The second consideration involves test reliability andthe results of the cluster analysis. Cluster analyses areonly as reliable as the measures on which subjects areclustered. This study demonstrated high reliabilitycoefficients for derived scales Sx and Os, in the main studyand in the validation study. Although no such coefficientwas available for the K (EDSS) scale (studies have examinedwhether two or more raters assign the same EDSS score to apatient, rather than whether raters’ overall assignmentscorrelate well), reasonable agreement has been shown betweenraters’ EDSS assignments (Francis et al. 1991). Despitedrawbacks of the EDSS (Willoughby and Paty 1988), it remainsthe best available method for assessing disability orimpairment in MS. The EDSS also has the advantage ofproviding a single numerical summary of an individual’sphysical problems. Another aspect of the cluster analysisregards the use of Ward’s method and its particular inherentbias. This method tends to identify spherically-shapedclusters, which in the current study refers to Severitygroups. The fact that salient—features profiles were alsoidentified, and validated, gives confidence that such groupsexist. However, because Ward’s method is not as sensitiveto outliers as are some cluster methods, it is possible thatsmall patient groups, showing different coping styles, maynot have been identified.117The third consideration involves validity, forexperimental inference (internal and external validity)(Campbell and Stanley 1966) and for the instruments used inthe study (construct, content, and criterion—relatedvalidity, with construct validity the most important) (Meehl1973; ICerlinger 1986; Messick 1989). In the realm ofexperimental inference, the study strove for high internalvalidity in terms of careful selection of patients, in thechoice of MRI sites most likely both to show variability andto underlie psychological change, and in the setting of astringent alpha for significance in the MRI lesion-locationanalysis. There can be a trade-off (ICerlinger 1986) betweeninternal and external validity, because internal controlswhich give confidence can also potentially limitgeneralizability. However, several arguments can be madefor generalizability of the results of this study to othersimilar MS populations. Selection of patients was done on abasis that would apply in other MS centers; for example, thecriteria for a diagnosis of CDMS (Poser et al. 1983) arewidely used. As well, the exclusion of certain anatomicalsites from the MRI analysis was not done for reasonsspecific to this sample. Of the 23 sites which wereexcluded, nine sites were excluded for theoretical reasons.Seven sites were excluded because more than 65% of patientshad lesions there; these sites were almost allperiventricular, and all seven sites were in regions whichhave been reported as high incidence in MS (Escourolle and118Poirier 1978; Valk and van der Knaap 1989). Another sevensites were excluded because fewer than 5% of patients hadlesions there; most of these sites were gray—matter, and allare known as low incidence in MS (Escourolle and Poirier1978; Valk and van der Knaap 1989).In the realm of the validity of instruments used, theconstruct validity of Sx and Ds are supported for thefollowing reasons: (1) items for Sx and Ds were chosen fromthe MMPI to satisfy theoretical constructs - the patient’sview of physical difficulties arising from MS (the Sx scale)and distress (the Ds scale). Judgments were madeindependently by the two raters, and agreement was requiredfor inclusion of any item on either scale. (2) the Sx andDs scales do not measure the same phenomena yet do have arelation to each other, as evidenced by their correlation of0.47. (3) patients in the Depressed group, who score highon Ds, also score high on the Beck Depression Inventory.(4) MS and controls scored significantly differently fromeach other on both Sx and Ds.The validity of the K (EDSS) scale as a measure of MSclinical problems must be assumed, not only because it isthe most widely used clinical evaluation tool in MS, butbecause this scale partly defines the diagnosis of MS.There is also divergent validity between Sx and K, andbetween K and Ds, pairs of measures which, as expected fromthe literature, do not correlate highly. The lowcorrelation between Sx and K is consistent with reports119that, as physicians are well aware, patient self—report doesnot always agree with results of clinical assessment (Taylor1991); a specific example from MS was demonstrated by Beattyand Monson (1991) in which many patients’ opinions of theirmemory difficulties were considerably different from resultsof objective memory tests. The lack of correlation betweenK and Ds is supported by previous evidence that there is nosignificant relation between degree of disability asmeasured by the EDSS, and the presence of depressivedisorder (Joffe et al. 1989).In the case of MRI, issues of validity are lessapplicable than in the case of the scales cited above inwhich abstract concepts are being measured. Nevertheless,MRI could be said to have validity for this study becauseits distinction between different tissues is based largelyon water—content, which changes in MS as a result of thebreakdown of myelin. The validity of MRI is supported bythe fact that this imaging process reveals the same lesionsas does CT (Bydder et al. 1982) as well as many morelesions, and by the fact that MRI data concur with those ofpost—mortem assessment. Stewart et al. (1984, 1986) haveshown that areas of increased signal intensity on MRIcorrespond to regions of MS pathology in post—mortem MSbrains. Although the correlation between MRI lesion sitesand physical problems is far from perfect (Isaac et al.1988; Paty et al. 1988), there is a relation betweenconfluence of lesions and course of disease, with chronic120progressive MS patients showing notably more confluence thanpatients with benign disease (Koopmans et al. 1989). Aswell, examination of MS patients using sensitiveneuropsychological tests has demonstrated that cerebrallesions may not, in fact, be silent and that total MRIlesion area can predict cognitive dysfunction (Rao et al.1989a)Regarding neuropsychological tests, the validity of theWAIS—R as a measure of global intelligence has been well—established, in comparisons of WAIS scores to other measuresof academic success and to alternative tests of intellectualfunctioning (Wechsler 1981). The WAIS (1955) and itssuccessor the WAIS-R (1981) have been used in thousands ofstudies to measure intelligence in its various aspects.Though originally standardized on normal individuals, theWAIS-R has been used extensively in clinical populations(Lezak 1983) including MS (Rao 1986) and can be valuable inassessment of such groups (Strub and Black 1985). Onepotential problem in the use of intelligence tests inneurological patients is that patients’ illnesses may leadthem to score badly for reasons not related to intelligence.However, this potential problem should not be a factor inthe current study, because tests with strong sensory ormotor components were not used in the analysis, and becausethe patients had relatively mild physical problems fromtheir illness.121Further to the issue of validity is that a validationstudy was done on the positive results of the main study,providing replication and thereby reducing threats to bothexperimental and instrumental validity. At the level ofinstruments used, Sx and Ds scales were again analysed inthe validation study for reliability (internal consistency),and for their ability to distinguish MS subjects fromcontrols in the new group of subjects used. At theexperimental level, the cluster analysis was againperformed, as were MRI and neuropsychological analyses whichshowed significance in the main study.A fourth methodological consideration in the currentstudy was the following. As in any anatomical localizationstudy of “experiments of nature,” this investigation wasconstrained by the practical need to delineate anatomicalregions as separate one from another. In this case, 50sites were pre—determined based on previous experience withMRI and MS. However, biological lesions do not always fitneatly into such logical schemes. Lesions frequentlyoverlap site boundaries. In other cases two or more lesionsmay be located in one site. Similarly, lesions frequentlyvary in volume, and may not be fully captured in the axialdirection.The fifth consideration is the relationship betweenwhite matter lesions and behavior. Current understandingsof the brain are based largely on the role of gray—matternuclei. Because white matter consists of tracts carrying122information between and among many nuclei, the effects ofwhite matter lesions on behavior are even more complex, andless understood, than similar effects of gray matterlesions. As an example, the (largely white matter) lesionsof MS frequently do not correspond to clinical signs (Patyet al. 1985; Stevens et al. 1986; Isaac et al. 1988;Koopmans et al. 1989), especially when lesions are above thebrainstem. Numerous studies have probed the consequences ofwhite matter lesions. Rao et al. (1989a) showed that totallesion area did predict cognitive dysfunction in MS, andthat smaller size of the corpus callosum predicted lowertest performance on measures of mental processing speed.Rao et al. (1989b) showed that corpus callosum atrophyaffected dichotic listening performance, suggestingdecreased efficiency of inter—hemispheric information flowin MS. Research into relations between white—matter changesand behavior is also being conducted in other disorders suchas Alzheimer’s Disease (Harrell et al. 1991). However, ourunderstanding of such relations is limited.Sixth, although the overall sample size was large,there was no way of ensuring that all groups to be analysedwould be large. In the clustering portion of the study, allprofile groups had 14 or more subjects. However, in thesubsequent MRI analysis, the ‘Depressed’ group (n=14)divided into two equal subgroups - those with the rightparietal lesion and those without that lesion. Therefore,two groups of seven subjects were used in the123neuropsychological analyses. In the validation study, the‘Depressed’ patients again divided into two subgroups: thosewith the lesion (n=4) and those without the lesion (n=6).Because these groups were small, outliers may have hadprofound effects on the statistical tests.The seventh consideration involves Type I and Type IIerror. The probability of Type I error (false rejection ofthe null hypothesis) in a significance test is controlled bythe setting of the alpha level. The levels of alpha set inthis study were guided by a desire to obtain results whichwere interpretable within the proposed model. In the MRIanalysis, for example, alpha was set at p 0.005, so thatany connection between profile and lesion site could beviewed with confidence, and so that a theory would not bedeveloped based on weak tests of significance. Hypotheticaloutcomes, discussed in Chapter Three and outlined inAppendix B, indicated that p 0.005 was a point at whichdata were becoming compelling and interpretable within themodel. Although many scientific studies employ an alphalevel of p < 0.05, it should be emphasized that this levelis arbitrary. The choice of alpha is at the discretion ofthe experimenter, and is based on the ramifications ofmaking either a Type I error or a Type II error (Hays 1988).For example, when a new drug, with potential side—effects,is being tested, one would want to see treatment effects ata relatively stringent (low) alpha before calling the drugeffective; on the other hand a relatively innocuous124treatment such as art therapy might be worth implementingeven if effects were significant at only a less stringent(higher) alpha. Regarding the chi-squared analyses of thecurrent study, the contrast—wise significance level waschosen as the level at which results would be meaningful.However, even though the contrast—wise error rate was set atp < 0.005, the experiment-wise error rate for the MRIanalysis was potentially high. Given that this study wasexploratory, it was deemed better to make an experiment—wiseerror than a Type II error, because there was no expense ordanger associated with a Type I error.However, another aspect of setting the alpha level isthe underlying scale of measurement, and expected variation.In the MRI data, except for small periventricular lesions,there is little or no variation in the normal sample, whichshould be virtually lesion—free. In contrast, in cognitivetest results, considerable variation is expected in thenormal sample. Therefore, alpha levels for theneuropsychological analysis were set at p 0.05 because ofthe need to detect subtle differences in test scores inwhich large inter-subject variability is present.The eighth and final issue of concern is experimentwise error, which is the increased chance of Type I errorwhenever numerous tests of significance are performed. Inthe current study, it is difficult to determine the level ofexperiment—wise error, due to the number of independentanalyses involving multiple measures. For the main study,125clearly experiment—wise error was high. Therefore, avalidation component was included to give a higher level ofconfidence in the results. Positive findings were thosewhich were replicated in the validation study. Initialpositive results which were not replicated remain ambiguous.Based on this procedure, the overall results of the studyare illustrated schematically in figure IX, which showsresults from the cluster analysis, MRI analysis andneuropsychological analysis. For each profile group,results in a given part of the study are shown as a plus orminus. So, if an analysis showed positive results in themain study which were replicated in the validation study,this is shown as C++); if a finding was positive in the mainstudy but not validated, it is shown as (+-). If ananalysis was not done in the validation study because theoriginal result was negative, it is represented as (nd).126FIGURE IXSchematic illustration of overall results(+) = positive results(—) = negative resultsC++) = positive results in both the main study and thevalidation study(nd) analysis not done‘Depression’ Denial Exag. Som. SeverityNeuropsychAnalysisClusterAnalysis++ ++ ++MRI ++Analysis++-nd -nd -nd-nd -nd -nd127FindingsThe cardinal finding of the current study supportsHypothesis one, that distinctive psychological profiles ofaccommodation exist and can be identified. Specifically,the finding involved the empirical identification andvalidation of four psychological profiles or copingresponses in patients with MS. These profiles areconsistent with qualitative descriptions of psychologicalresponses in MS patients, namely ‘Depression’, Denial,Exaggerated Somatic and Severity.The ‘Depressed’ group showed high distress relative todisability; frequency of this profile was 14/99 (14%) in themain study and 10/43 (23%) in the validation study. TheDenial group showed low scores on both self—rating scalesrelative to their higher scores on K (EDSS); this profileoccurred in 32/99 (32%) patients in the main study and 12/43(28%) in the validation study. The Exaggerated Somaticgroup showed high scores on Sx relative to scores on K; thisprofile occurred in 22/99 (22%) patients in the main studyand 12/43 (28%) patients on validation. The Severity groupshowed similar scores on all three measures, whether low orhigh; this profile occurred in 31/99 (31%) patients in themain study and 9/43 (21%) on validation.The second finding supports Hypothesis two, thatmembership in psychological groups is not related to extentof pathology as measured by number of sites with lesions.128Data from the study show that psychological groups are notdistinguished by number of sites with lesions, suggestingthat membership in groups is not a function of amount ofbiological disease.Hypothesis three, that membership in profile groups wasrelated to location of lesions, was not supported for threeof the four profiles. That is, no relations were found forthe Denial, Exaggerated Somatic and Severity groups, andsite of lesion. However, for a subgroup of the ‘Depressed’profile, a significant relation was found. Specifically,seven out of 14 of the ‘Depressed’ subjects (50%) in themain study, and four out of 10 subjects (40%) in thevalidation study had lesion(s) in the gray matter-whitematter junction of the right parietal lobe. Therefore, therelation was found in both the main study and the validationstudy, in 11/142 patients (7.7%). A relation between‘Depression’ and and right—hemisphere lesions accords withsome past reports on gray matter (GM) lesions, becauserelations have been suggested between right—hemispherelesions and depression (Folstein et al. 1977; Lezak 1983).However, the bulk of the neurology literature has associatedleft—hemisphere lesions with depression (Robinson and Price1982, Lezak 1983), and right-hemisphere lesions with denialand anosagnosia (Lezak 1983). Nevertheless, any apparentdiscrepancy between the finding of this study, and the bulkof the literature on brain lesions, emphasizes how little isknown about the behavioural consequences of white matter129lesions. There is no reason to believe that white matterlesions in the parietal lobe, affecting tracts which may betransferring information from, for example, the occipitallobe to the frontal lobe, will affect functions classicallyassociated with the parietal lobe.Hypothesis four of the study, that group membershipwould be related to performance on cognitive tests, was notsupported for most MS patients. That is, for the majorityof MS groups, no relation was found between group membershipand specific cognitive impairment. However, there was astrong indication in the main study that the ‘Depressed’group with the right parietal lesion had widespreadcognitive impairment, because this group scored lowest ofall groups on 18/19 neuropsychological tests. Thevalidation study did not confirm this cognitive impairment,so that the finding remains unclear. The potentiallyimportant symptom triad of ‘Depression’, a right parietallesion, and general cognitive dysfunction, neverthelesswarrants further study. As discussed earlier, the smallsize of the ‘Depressed’ group with the lesion in thevalidation study (n=4), could have precluded findings ofsignificance.130ImplicationsThe first implications to be discussed concern theprofiles illustrated in the first vertical line of figureIX. Clearly, the profiles were not defined by K (EDSS)scores, as several pairs of clusters had similar K scoresyet fell into separate profile groups based on differencesin Sx and Ds scores. This fact underscores a central themeof this study, that amount of clinical disease (level ofdisability or impairment) does not alone determinepsychological response. Therefore, not only could an MSpatient display any of a number of different responses, butany one response is not necessarily characteristic of acertain stage of clinical disease.These data on profiles do not support the position thatthere is an “MS personality”. Rather these unique profilessuggest an interaction between pre—morbid personality orcoping style and the diagnosis of MS. The only possibleexception in the current study is the case of the 11subjects with ‘Depression’ and a lesion at the gray matter—white matter junction of the right parietal lobe.The identification of separate groups also supports theconcept that individuals faced with life crises do not allrespond similarly. Heterogeneity of response to illness hasbeen demonstrated in rheumatoid arthritis (McFarlane et al.1987; Keefe et al. 1989), head injury (Nockleby and Deaton1987; Moore et al. 1989), spinal cord injury (Frank et al.1311987), coronary bypass surgery (Clark and Kionoff 1988), andother illnesses as well as in non—medical crises (Kessler etal. 1985) . Moreover, the coping mechanisms suggested inthese studies are similar to those reported in otherdiseases (Kessler et al. 1985). Therefore, in general, onewould propose an interactive model based on adjustmentstrategy prior to diagnosis, coupled with increased stress.Such a model is further buttressed by the MRI datasummarized on the second line of figure IX. For the MSsubjects, particularly in the Denial, Exaggerated Somaticand Severity profiles, no relation was found between profilemembership, and site of lesions. This study thereforetested the possibility of an organic basis to patients’responses, and did not support such an organic model. Onepossible explanation must be that the majority of MSpatients with mild physical problems who experiencepsychological difficulties do so as a reaction to the stressof MS rather than as a direct result of lesions.The exception to this model was the ‘Depressed’ group.For 11 of 24 of these subjects, ‘Depression’ was associatedwith a certain lesion in the right parietal lobe.Therefore, for the individual subject with ‘Depression’,there may or may not be an organic basis.The lack of significant relations between profiles andlesion sites need not contradict previous reports thatpsychological change such as euphoria may be organically—based (Rabins et al. 1986; Minden and Schiffer 1990a).132Though denial and euphoria share features in common, it maybe that denial in patients with mild physical problems ismost often a reactive response, while euphoria in moreseverely disabled patients is the result of disruption ofbrain function. Given the selective sampling procedures,questions about severely disabled patients could not bedirectly addressed in this study. However, for mild MSpatients, the denial response appears to be reactive.For the neuropsychological data summarized in the thirdline of figure IX, it was predicted that the Denial profilegroup would do poorly on tests requiring insight. Thisprediction was not borne out; Deniers scored lower thanmost, but not all, other groups on Similarities andcategories tests. Similarly, it was predicted that‘Depressed’ groups would do poorly on tests requiring newlearning and memory. This prediction was also not clearlysupported, as there were no memory tests on which the twogroups of ‘Depressed’ patients scored significantly lowerthan did other groups.Neuropsychological data from this study are relevant tothe general issue of the relation between depression andcognitive impairment. While one group (n=7) of ‘Depressed’patients in the main study showed significant cognitiveimpairment, the other group (n=7) did not. In fact, thelatter group (‘Depressed’ but without the characteristiclesion) performed better than normal controls on severaltests, while the group with the lesion performed worse than133the other groups. There is a considerable literature on thelink between depression and cognitive impairment (reviewedin Weingartner and Silberman 1982). In theory, causation(if such exists) could run in either direction, withdepression causing cognitive dysfunction or cognitivedysfunction causing depression. In contrast, theperformance of ‘Depressed’ patients in the current studyshows that depression, as defined here, need not necessarilycause cognitive impairment. This contention is consistentwith the work of others who have concluded that depressionand cognitive dysfunction can be independent in MS (DePauloand Foistein 1978; Peyser et al. 1980a).This study was not designed to examine or evaluatetherapies for psychological problems. However, severalspeculative statements can be made. Patients whosepsychological response is problematic, and for whom there isevidence of a biological basis to the response, may bestrong candidates for treatment of the MS disease processitself. Patients whose psychological problems have noidentifiable biological basis may be strong candidates forpsychiatric treatment, either medication or psychotherapy ora combination thereof. Either group of patients (likelyorganic, and likely—reactive) can, however, be consideredfor both anti-MS and psychiatric therapy.In the realm of psychiatric treatment, there isevidence that both chemotherapy and psychotherapy can beuseful to MS patients. Schiffer and Wineinan (1990) showed134that the anti-depressant medication desipramine has a modestbeneficial effect in serious depression associated with MS.Several research groups have shown benefits frompsychotherapy for MS patients (Larcombe and Wilson 1984;Schiffer 1987). Crawford and Mclvor (1985) demonstratedthat group psychotherapy for MS patients resulted indecreased depression and anxiety, and increased self—conceptand self-direction. Frank et al. (1987) have suggested thatwell-designed psychotherapeutic treatment may assistpatients to forge their own most effective attitudes towardthe disease. Several authors add that the existence ofpsychological problems which apparently have an organicbasis are not a contraindication to psychotherapy (Talbottet al. 1988; Minden and Noes 1990).Further research may clarify the existence of asubgroup of organically ‘Depressed’ patients, and possibletreatment implications. For now, while this study will beof interest to clinicians, its data are too preliminary toform the basis for a recommendation of regular testing atthe MS Clinic.Further CommentsAlthough four profiles were found, there is still thequestion of whether other, different, profiles might alsoexist. Though the existence of alternative profiles istheoretically possible, any such groups would have contained135fewer than five subjects, the smallest cluster size.Therefore, error in failing to identify such clusters wouldhave been small. As well, the many reports in theliterature which have commented on psychological profiles inMS do not suggest notable alternatives to the four profilesidentified here.The study has not attempted to place value judgments onany or all psychological responses to MS, or to say that oneresponse is more appropriate than another. All responsesare appropriate in the view of individuals who show them.Although it is possible that one response might havedifferent long—term effects on the course of the illnessthan another, if, for example, relapsing disease is relatedto perceived stress, such an assessment would require along—term study. However, the coping response in allprobability does affect accommodation to diagnosis andsubsequent enjoyment of life.The ‘Depression’ in this study was defined as anincrease in the level of distress relative to theindividual’s level of physical difficulty from the disease.This definition was thought to be appropriate, because onemight expect a general increase in distress in MS patients.The ‘Depression’ of this study, then, was not the same asthe depression defined by the psychiatric Diagnostic andStatistical Manual of Mental Disorders (DSM-IIIR). In theDSM-IIIR, criteria for diagnosis of major depression are (1)depressed mood, (2) markedly diminished interest in normal136activities, (3) significant weight loss or gain, (4)hypersomnia or insomnia, (5) psychomotor agitation orretardation, (6) fatigue or loss of energy, (7) feelings ofworthlessness or excessive guilt, (8) diminished ability tothink or concentrate, or indecisiveness, (9) recurrentthoughts of death; plans for suicide. The Ds scale of thisstudy, an important component of the definition of‘Depression’, contains items referring to all nine of theseDSM-IIIR criteria. Therefore, although the ‘Depression’ ofthis study does not use the criteria of psychiatricdepression, there is expected to be considerable overlapbetween the two definitions.The division of ‘Depressed’ patients into two subgroupssuggests the terms ‘reactive’ and ‘endogenous’, which havebeen used to describe, respectively, depression with a clearprecipitating stressor, and depression which may insteadhave an organic basis. Clearly, the two categories are endsof a continuum, and do overlap (Kaplan and Sadock 1991).However, in a general sense it may be that ‘Depressed’patients with the parietal lesion are displaying anendogenous depression, while ‘Depressed’ patients withoutthe lesion are displaying a reactive depression. For thosewith reactive depression, the stressor is the diagnosis ofMS. For those with endogenous depression, their lesion mayhave contributed to their mood. This distinction does not,however, exclude either group from any possible therapy, nordoes it specify treatment. For example, although137psychiatric patients with apparently endogenous depressionare considered good candidates for anti—depressantmedication or electro—convulsive therapy, one cannot assumethat such therapies will be effective for these MS patients,because the cause of their depression may be different.Nevertheless, a distinction between reactive and endogenousdepression is useful. Such a distinction recalls thecentral point of the study, and a central point inpsychiatry- the quest to know whether a patient’spsychological problems are, or are not, organically—based.This study emphasized the importance of patients’personal views of their illness, which form the basis of thetwo self—report measures of the three measures used in thecluster analysis. A patient’s belief about his/her level ofillness can differ considerably from the physician’s belief,as illustrated in this study in the differences between Kand Sx for many patients. Beatty and Monson (1991), in astudy on memory problems in MS, have also shown thatpatients’ reports can differ markedly from those of theirphysicians. In the present study, there is a strongercorrelation between Sx and Ds (0.47; p < 0.001) than betweenK and Ds (0.10, p < 0.15), demonstrating that patients’views of reality influence their mood considerably more thando objective measures of reality, Clearly, it isenlightening to consider a patient’s perception of theillness.138Profiles identified in this study represent copingstrategies, although this study did not attempt to examineall possible methods of coping, which are cognitive andbehavioural efforts to master, tolerate or reduce demandswhich tax a person’s resources (Lazarus and Folkman 1984).The term “coping” is used broadly in the literature toinclude not only attitudes and behaviours towards illnesssuch as the psychological profiles of this study, but alsoexplanatory strategies such as those involving religiousconviction (Matson and Brooks 1977). Many coping methodscould not be examined in this study, the purpose of whichwas to cite from the literature, and empirically identify,responses common in MS.Data from this study add to those of investigators whohave questioned the widespread application of stage theoriesto models of coping. Stage theories assume that individualsprogress through predictable levels during, or following,important life processes or events. Such theories have beeninfluential in psychology, for descriptions of cognitivedevelopment in children (Piaget 1929), moral maturation(Kohlberg 1958), responses to death and dying (Kubler-Ross1969), and progress through male adulthood (Levinson 1978).However, several recent authors (Kessler et al. 1985; Franket al. 1987) have commented that such models are notstrongly supported by empirical evidence. Because stagemodels frequently imply that patients must work through aseries of attitudes before coming to accept their139situations, Frank et al. (1987) also say that stage modelshave encouraged passivity on the part of clinicians, andhave retarded the development of modes of treatment whichwould help patients control their own responses.The stage concept was applied to MS by Matson andBrooks (1977) who proposed a model in which patients weresaid to pass through four stages of coping with theirillness. It was suggested that MS patients more recentlydiagnosed behaved differently toward their illness than didpatients diagnosed long before. Similarly, Jouvent et al.(1989) reported that depression was more common in patientswho had lived with the disease for less than two years, thanin patients who had lived with MS for more than six years.Data from the current study show that psychological profilegroups differ slightly on number of years since onset offirst symptoms, but do not differ significantly on number ofyears since diagnosis. Data from the current study, then,do not clearly support a stage model based on length of timean individual has been living with the disease. Age data donot clearly support a stage model either, because the Denialgroup was oldest in the main study, but the Severity groupwas oldest in the validation study.Data from the current study do not, however, show thatpsychological response is often a function of site—of—lesionin patients with early-stage MS. What other models, then,might be considered to explain different psychologicalresponses? As mentioned, several studies suggest that140depression is most common among patients recently diagnosed.But because more—recently diagnosed patients are usuallyalso younger than patients less—recently diagnosed, it maybe that patients respond differently based on their “stageof life”. Such a model could be investigated further.However, while there were significant differences betweenages of profile groups in this study, the fact that allgroups had mean ages in their 30s suggests that groups werenot at markedly different stages of maturity.Data from this study do not support a stage model basedon either level of clinical disease as measured by EDSSscore, or level of biological disease as measured by numberof sites with lesions. It is nevertheless theoreticallypossible that patients might change their response at somepoint based not on clinical or biological stage, but forother, psychological, reasons. That is, a patient may usedenial for a time, find it inadequate, and collapse intodepression. An alternative possible model to explain theexistence of various profiles is that of individualdifferences in dimensions of personality, which lead to avariety of ways of coping with stress.141Contributions of the Study;Suggestions for Future ResearchThis study has made contributions in both evidence andmethodology. In the realm of evidence, the study hasempirically validated the existence of a heterogeneity ofcoping styles in MS. These data should caution researchersnot to expect any one psychological outcome in MS. Datafrom this study also suggest that different profiles may notbe the result of stage of biological disease. These datacould encourage more active psychotherapeutic interventionfor MS patients, on the basis that patients need nottraverse an inevitable series of psychological crises indealing with their MS.In the larger health field, support for heterogeneityof response to a given situation fits with an importantissue in medicine and psychology, and in the relatively newfield of health psychology. It has been observed innumerous kinds of crises that individuals responddifferently one from another. So, for example, when twoindividuals are both rendered paraplegic, one may sink intodespair and dependence, while another may become more activeand seek independence. This phenomenon occurs, then, in anumber of different situations, even in ones in which thebrain has not been affected. Data from the current studycannot address the larger issue of reasons for differentresponses, but the study does remind researchers and142clinicians not to expect certain physical or psychologicaloutcomes in individuals based on a single agent or event.In the realm of methodology, the study avoided many ofthe problems for which past psychological studies in MS havebeen criticised (VanderPlate 1984; Rao 1986; Minden andSchiffer 1990; Peyser et al. 1990). The study had a largesample, chosen from a clinic population of more than 2,000patients. The sample was well—characterized and fullydescribed. The study tested only relapsing/remittingpatients, and controlled for the cyclic nature of thedisease by testing solely individuals in remission. Also inthe realm of methodology, the study provided furtherevidence of the utility of cluster analysis for clarifyingconcepts involving several variables.There are numerous questions which could be asked infuture work, including: (1) Are various therapies moreeffective in MS patients if it is first determined whetherthere might be an organic basis to any psychological change?(2) For ‘Depressed’ patients with the characteristic lesionin the right parietal lobe, does cognitive dysfunction oftenarise? (3) Do different coping responses have differenteffects on the course of the illness?143APPENDIX ACORRELATIONAL STUDIES WHICH HAVE BEEN DONE IN MS, BETWEENPSYCHOLOGICAL VARIABLES AND PHYSICAL/DISEASE VARIABLESA. Affective Problems and Extent of Disabilitya) Studies Suggesting Correlations(1) Cleeland et al. (1970) showed that depression, asgauged by the NNPI, is greater for patients in exacerbationthan those in remission.(2) Mclvor et al. (1984) showed that depression, asgauged by the Beck Depression Inventory (BDI), is greater inmore-disabled spinal MS patients than in less-disabled ones.It should be noted that both the BDI and the Depressionscale of the MIVIPI contain items which could be endorsedpurely on the basis of physical symptoms of MS.b) Studies Suggesting Little or No Correlation(1) Joffe et al. (1989) found no direct relationshipbetween degree of disability and mood disorders. Patientswith major depression were generally less disabled thanthose without such psychiatric diagnosis.(2) Minden et al. (1989) reported no relationshipbetween severity of disability and depression as measured bythe BDI.144B. Cognitive Function and Extent of Disabilitya) Studies Suggesting Correlations(1) Beatty and Gange (1977) suggested a relationbetween memory dysfunction and motor dysfunction, showingthat correlations between motor and memory performance wereconsistently higher in MS patients than in controls.(2) Surridge (1969) showed a significant associationbetween intellectual deterioration and physical disability.(3) Stenager et al. (1989b) suggested that aspects ofcognition including verbal and visual memory decline withincreasing score on the Expanded Disability Status Scale(EDSS), a measure of disability/impairment for MS (Kurtzke1983)b) Studies Suggesting Little or No Correlation(1) Van den Burg et al. (1987) showed no relationbetween most intellectual abilities and MS progression, andonly weak relations between memory measures and the disease.(2) Rao et al. (1984) reported no relation betweendegree of memory disturbance and EDSS score.(3) Lyon-Caen et al. (1986) found no correlationbetween cognitive difficulties, and degree of illness asmeasured by EDSS.C. Affective Problems and MRI Lesion Measuresa) Studies Suggesting Correlations(1) Honer et al. (1987) reported that MS patients withpsychiatric disorders had more temporal—lobe lesions than145patients with no such disorders. However, the authorsreported no relation between total lesion burden and thepresence of psychiatric disorder in MS patients, whichsuggests that lesion site (rather than amount) may becritical.(2) Rabins et al. (1986) showed that MS patients withbrain involvement had more depression than MS patients withonly spinal—cord involvement.b) Studies Suggesting Little or No Correlation(1) Logsdail et al. (1988) and Ron and Logsdail (1989)reported no significant associations between total MRIlesion scores and severity of psychiatric problems.D. Cognitive Problems and MRI Lesion Measuresa) Studies Suggesting Correlations(1) Callanan et al. (1989) reported that IQ deficit anddefective auditory attention were significantly correlatedwith the degree of brain pathology detected by MRI.(2) Franklin et al. (1988) showed a relation betweencognitive impairment and total brain—lesion area.(3) Rao et al. (l989a) showed that several MRImeasures, particularly total lesion area, were significantlydifferent in cognitively impaired patients from patients notimpaired.b) Studies Suggesting Little or No Correlation(1) Huber et al. (1987) reported no correlation betweenamount of dementia in MS patients and any of the following146MRI measures: number of lesions, size of lesions, and extentof generalized cerebral atrophy. However, the authorsreported a modest correlation between dementia and corpuscallosum atrophy.147MRI LESION-LOCATION ANALYSIS: HYPOTHETICAL DISTRIBUTIONS OFSUBJECTS WITHIN GROUPS, AND ASSOCIATED P-VALUESAppendix B.l. With zero noise in 3/4 profile groups, how manymembers of a profile group must have lesions for significance?(# = number of members w/lesion) (asterisk=p 0.005)n14 n=22 n32# matrix p-value123456781 0 0 0 .10513 32 22 312 0 0 0 .006212 32 22 31*3 0 0 0 .000311 32 22 31matrix p—value001014 32 21 310 0 2 0 .06714 32 20 310 0 3 0 .012714 32 19 31*0 0 4 0 .002214 32 18 31*0 0 5 0 .000414 32 17 31matrix p—value010014 31 22 310 2 0 0 .23314 30 22 310 3 0 0 .09114 29 22 310 4 0 0 .033114 28 22 310 5 0 0 .011614 27 22 31*0 6 0 0 .003914 26 22 31*0 7 0 0 .001314 25 22 31*0 8 0 0 .000414 24 22 31APPENDIX B148Appendix B.2. With minimum noise in 3/4 profile groups, how manymembers of a group must have lesions for significance?n14 n=22 n32# matrix p-value matrix p—value matrix p—value1236789103 1 1 1 .07811 31 21 301 1 3 1 .36613 31 19 301 1 4 1 .13313 31 18 301 9 1 1 .0093231 10 1 1 .003322**4 4 1 1 1 .009210 31 21 30*5 5 1 1 1 .00009 31 21 301 1 5 1 .038 1 5 1 1 .28117 13 27 21 301 1 6 1 .0092 1 6 1 1 .13816 26*1 1 7 1 .0019 1 7 1 1 .06115 25*1 1 8 1 .0004 1 8 1 1 .02514 24121 11 1 1 .001121*1 12 1 1 .000320149Appendix B.3. With a large amount of noise (50%) in 3/4 groups,how many members of a group must have lesions for statisticalsignificance?n14 n=22 n=32matrix p—value14 16 11 160 16 11 15matrix p—value1 16 22 16 .00057 16 0 157116118116 .078matrix p-value7 32 11 160.0073* *.000013 16 11 16 .0326112 16 11 16 .11032* *7 16 21 16 .0024 7 31 11 16 .00011 1*7 16 20 16 .0091 7 30 11 16 .00052 2*7 16 19 16 .0291 7 29 11 16 .00193 3I lii7 28 11 164.00587 27 11 16 .015657 26 11 16 .037467 25 11 16 .08007150APPENDIX CMMPI ITEMS COMPRISING Sx AND Ds SCALESSx scale:007: My hands and feet are usually warm enough.009: I am about as able to work as I ever was.018: I am very seldom troubled by constipation.023: I am troubled by attacks of nausea and vomiting.036: I seldom worry about my health.047: Once a week or oftener I feel suddenly hot all over,without apparent cause.051: I am in just as good physical health as most of myfriends.061: Parts of my body often have feelings like burning,tingling, crawling, or like ‘going to sleep’.062: I have had no difficulty in starting or holding mybowel movement.067: I hardly ever feel pain in the back of my neck.100: I have little or no trouble with my muscles twitchingor jumping.116: My speech is the same as always (not faster or sloweror slurring: no hoarseness.)150: During the past few years I have been well most of thetime.157: I have never felt better in my life than I do now.160: I do not tire quickly.172: I seldom or never have dizzy spells.151179: I am afraid of losing my mind.182: My hearing is apparently as good as that of mostpeople.183: I frequently notice my hand shakes when I try to dosomething.184: My hands have not become clumsy or awkward.185: I can read a long while without tiring my eyes.186: I feel weak all over much of the time.189: I have had no difficulty in keeping my balance inwalking.191: I have had attacks in which I could not control mymovements or speech but in which I knew what was goingon around me.239: I have few or no pains.268: I have numbness in one or more regions of my skin.269: My eyesight is as good as it has been for years.276: I do not often notice my ears ringing or buzzing.313: I have never been paralyzed or had any unusual weaknessof any of my muscles.314: Sometimes my voice leaves me or changes even though Ihave no cold.393: I have had no difficulty starting or holding my urine.152Ds scale:002: I have a good appetite.003: I wake up fresh and rested most mornings.008: My daily life is full of things that keep meinterested.010: There seems to be a lump in my throat much of the time.016: I am sure I get a raw deal from life.022: At times I have fits of laughing and crying that Icannot control.040: Most any time I would rather sit and daydream than todo anything else.041: I have had periods of days, weeks or months when Icouldn’t take care of things because I couldn’t getgoing.043: My sleep is fitful and disturbed.046: My judgeinent is better than it ever was.066: I wish I could be as happy as others seem to be.075: Most of the time I feel blue.083: These days I find it hard not to give up hope ofamounting to something.085: I am certainly lacking in self-confidence.087: I usually feel that life is worthwhile.101: I don’t seem to care what happens to me.104: I am happy most of the time.139: I certainly feel useless at times.149: Most nights I go to sleep without thoughts or ideasbothering me.153155: I cry easily.207: I can sleep during the day but not at night.213: I frequently find myself worrying about something.226: I hardly ever notice my heart pounding and I am seldomshort of breath.232: I brood a great deal.234: I have periods of such great restlessness that I cannotsit long in a chair.238: I believe I am no more nervous than most others.247: No one cares much what happens to you.252: I usually expect to succeed in things I do.294: Life is a strain for me much of the time.297: Even when I am with people I feel lonely much of thetime.300: I seem to make friends about as quickly as others do.318: I feel anxiety about something or someone almost allthe time.320: Most of the time I wish I were dead.340: Sometimes some unimportant thought will run through mymind and bother me for days.347: I am not unusually self—conscious.351: I very seldom have spells of the blues.354: People often disappoint me.356: My plans have frequently seemed so full of difficultiesthat I have had to give them up359: I have sometimes felt that difficulties were piling upso high that I could not overcome them.154364: I am usually calm and not easily upset.368: I am apt to take disappointments so keenly that I can’tput them out of my mind.370: At times I think I am no good at all.404: I feel like giving up quickly when things go wrong.408: I must admit that I have at times been worried beyondreason over something that really did not matter.421: The future seems hopeless to me.433: I sometimes feel that I am about to go to pieces.439: In the past 12 months, I have seriously thought abouttaking my life as a solution to personal problems.155REFERENCESAring, C.D. (1965) Observations on multiple sclerosis andconversion hysteria. Brain, 88, 663-674.Baldwin, N.y. (1952) A clinico-experimental investigationinto the psychologic aspects of multiple sclerosis.J Nerv Ment Dis, 115, 299-342.Baretz, R.M., and Stephenson, G.R. (1981) Emotionalresponses to multiple sclerosis. Psychosomatics, 22(2),117—127.Barnard, R.O., Triggs, M. (1974) Corpus callosum inmultiple sclerosis. J Neurol Neurosurg Psychiatry, 37,1259—1264.Beatty, P.A., and Gange, J.J. (1977) Neuropsychologicalaspects of multiple sclerosis. J Nerv Ment Dis, 164,42—50.Beatty, W.W., Goodkin, D.E., Monson, N., Beatty, P.A.,Hertsgaard, D. (1988) Anterograde and retrogradeamnesia in patients with chronic progressive multiplesclerosis. Arch Neurol, 45, 611-619.Beatty, W.W., Goodkin, D.E., Monson, N., Beatty, P.A.(1989) Cognitive disturbances in patients withrelapsing remitting multiple sclerosis. Arch Neurol,46, 1113—1119.Beatty, W.W., Monson, N. (1991) Metamemory in multiplesclerosis. J Clin Exp Neuropsychol, 13(2), 309—327.Beck, A.T., Beamesderfer, A. (1974) Assessment ofdepression: The depression inventory. In P. Pichot,ed., Psychological Measurements in Psychopharmacology.Modern Problems in Pharmacopsychiatry, vol. 7, pp. 151169, Basel: S. Karger.Benton, A.L. (1968) Differential behavioural effects infrontal lobe disease. Neuropsychologia, 6, 53—60.Bernstein, I.H. (1988) Applied Multivariate Analysis.New York: Springer—Verlag.Berrios, G.E., Quemada, J.I. (1990) Depressive illness inmultiple sclerosis. Clinical and theoretical aspects ofthe association. Br J Psychiatry, 156, 10-16.Blashfield, R.K. (1976) Mixture model tests of clusteranalysis: Accuracy of four agglomerative hierarchicalmethods. Psychol Bull, 83(3), 377—388.156Borberg, N.C., Zahle, V. (1946) on the psychopathology ofdisseminated sclerosis. Acta Psych Neurol, 21, 75-89.Bourestom, N.C., Howard, M.T. (1965) Personalitycharacteristics of three disability groups. Arch PhysMed Rehabil, 46(9), 626—632.Braceland, F.J., Giff in, M.E. (1950) The mental changesassociated with multiple sclerosis (an interim report).Proc Assc Res Nerv Ment Dis, 28, 450-455.Braffman, B.H., Zimmerman, R.A., Trojanowski, J.Q., Gonatas,N.K., Hickey, W.F., Schlaepfer, W.W. (1988) Brain MR:Pathologic correlation with gross and histopathology.Am J Neuroradiology, 9, 629—636.Brain, W.R. (1930) Critical review: Disseminated sclerosis.Q J Med, 23, 343—391.Bramwell, B. (1917) The prognosis in disseminatedsclerosis: duration in two hundred cases ofdisseminated sclerosis. Edinburgh J Med, 18, 16-23.Brown, S., , T. K. (1922) The mental symptoms ofmultiple sclerosis. Arch Neurol Psychiatry, 7, 629-634.Buzzard, T. (1897) Insular sclerosis and hysteria.Lancet, 1-4.Bydder, G.M., Steiner, R.E., Young, I.R., Hall, A.S.,Thomas, D.J., Marshall, J., Pallis, C.A., Legg, N.J.(1982) Clinical NMR imaging of the brain:l40 cases.AJR, 139, 215—236.Caine, E.D., Bamford, K.A., Schiffer,R.B., Shoulson,I.,Levy, S. (1986) A controlled neuropsychologicalcomparison of Huntington’s Disease and multiplesclerosis. Arch Neurol, 43, 249—254.Callanan, M.M., Logsdail, S.J., Ron, M.A., Warrington, E.K.(1989) Cognitive impairment in patients with clinicallyisolated lesions of the type seen in multiplesclerosis. Brain, 361—374.Campbell, D.T., Stanley, J.C. (1966) Experimental andqsi—experimental designs for research. Chicago: RandMcNally.Canter, A.H. (1951a) MMPI profiles in multiple sclerosis.J Consult Psychol, 15, 253—256.Canter, A.H. (195lb) Direct and indirect measures ofpsychological deficit in multiple sclerosis: part I.J Gen Psychol, 44, 3-25.157Charcot, J.M. (1877) Disseminated sclerosis: itssymptomatology. Lectures on the diseases of the nervoussystem, London: The New Sydenham Society.Churchland, P.S. (1986) Neurophilosophy: toward a unifiedscience of the mind/brain. Cambridge, Mass.: MIT Press.Clark, C., Klonoff, H. (1988) Empirically derived MMPIprofiles: Coronary bypass surgery. J Nerv Ment Dis,176(2), 101—106.Cleeland, C.S., Matthews, C.G., Hopper, C.L. (1970) MMPIprofiles in exacerbation and remission of multiplesclerosis. Psychol Rep, 27, 373—374.Cormack, R.M. (1971) A review of classification. J RoyalStat Soc, 134, 321—353.Cottrell, S.S., Wilson, S.A.K. (1926) The affectivesymptomatology of disseminated sclerosis. J NeurolPsychopath, 7(25), 1—30.Crawford, J.D., Mclvor, G.P. (1985) Group psychotherapy:Benefits in multiple sclerosis. Arch Phys Med Rehab,46, 810—813.Cronbach, L.J. (1951) Coefficient alpha and the internalstructure of tests. Psychometrika, 16(3), 297—334.Dabs, N.P., Rabins, P.V., Brooks, B.R., O’Donnell, P.(1983) Disease activity and emotional state in multiplesclerosis. Ann Neurol, 13, 573—577.DePaubo, J.R., Foistein, M.F. (1978) Psychiatricdisturbances in neurological patients: detection,recognition and hospital course. Ann Neurol, 4, 225—228.Dercum, F.X. (1912) A case of multiple cerebrospinalsclerosis, presenting unusual symptoms suggestingparesis. JAMA, 59(18), 1612—1614.Diagnostic and Statistical Manual of Mental Disorders (ThirdEdition - Revised) (DSM-IIIR) (1987) Washington DC:American Psychiatric Association.Escourelle, R., Poirier, J. (1978) Manual of BasicNeuropathology, 2nd ed. Philadelphia: W.B. Saunders Co.Fazekas, F., Of fenbacher, H., Fuchs, S., Schmidt, R.,Niederkorn, K., Homer, S., Lechner, H. (1988)Criteria for an increased specificity of MRIinterpretation in elderly subjects with suspectedmultiple sclerosis. Neurology, 38, 1822—1825.158Filskov, S.B., Grimm, B.H., Lewis, J.A. (1981) Brain-Behavior Relationships. In S.B. Filskov, T.J. Boll,eds., Handbook of Clinical Neuropsychology. vol.1, pp.39—73.Firth, D. (1948) The case of Augustus d’Este. Cambridge:Cambridge University Press.Fischer, J.S. (1989) Objective memory testing in multiplesclerosis. In K. Jensen et al, eds., Mental Disordersand Cognitive Deficits in multiple sclerosis. pp.39-49,London: John Libbey.Foistein, M.F., Maiberger, R., McHugh, P.R. (1977) Mooddisorder as a specific complication of stroke. J NeurolNeurosurg Psychiatry, 40, 1018-1020.Frank, R.G., Uxnlauf, R.L., Wonderlich, S.A., Askanazi, G.S.,Buckelew, S.P., Elliott, T.R. (1987) Differences incoping styles among persons with spinal cord injury: acluster-analytic approach. J Consult Clin Psychol,55(5), 727—731.Franklin, G.M., Heaton, R.K., Nelson, L.M., Filley, C.M.,Seibert, C. (1988) Correlation of neuropsychologicaland MRI findings in chronic/progressive multiplesclerosis. Neurology, 38, 1826—1829.Franklin, G.M., Nelson, L.M., Filley, C.M., Heaton, R.K.(1989) Cognitive loss in multiple sclerosis. ArchNeurol, 46, 162—167.Francis, D.A., Bain, P., Swan, A.V., Hughes, R.A.C. (1991)An assessment of disability rating scales used inmultiple sclerosis. Arch Neurol, 48, 299-301.Freedman, A.M., Kaplan, H.I., Sadock, B.J. (1976) ModernSynopsis of Comprehensive Textbook of Psychiatry/Il,2nd ed., Baltimore: Williams & Wilkins Co.George, A.E., de Leon, M.J., Kalnin, A., Rosner, L.,Goodgold, A., Chase, N. (1986) Leukoencephalopathy innormal and pathologic aging: MRI of brain lucencies, AmJ Neuroradiology, 7, 567—570.Gerard, G., Weisberg, L.A. (1986) MRI periventricularlesions in adults, Neurology, 36, 998—1001.Gilberstadt, H., Farkas, E. (1961) Another look at MNPIprofile types in multiple sclerosis. J Consult Psychol,5, 440—444.159Glass, G.V., Hopkins, K.D. (1984) Statistical Methods inEducation and Psychology, 2nd ed. Englewood Cliffs,N.J.: Prentice-Hall Inc.Golden, R.R., and Meehl, P.E. (1980) Detection ofbiological sex: an empirical test of cluster methods.Multivar Behav Res, 15, 475-496.Goldstein, G., Shelly, C.H. (1974) Neuropsychologicaldiagnosis of multiple sclerosis in a neuropsychiatricsetting. J Nerv Ment Dis, 158(4), 280—290.Goodstein, R.K., Ferrell, R.B. (1977) Multiple sclerosis —presenting as depressive illness. Dis Nerv Syst, 38,127—131.Gowers, W.R. (1896) A Manual of Diseases of the NervousSystem. Philadelphia: Blakiston Son & Co.Grant, I., McDonald, W.I., Trimble, M.R., Smith, E., Reed,R. (1984) Deficient learning and memory in early andmiddle phases of multiple sclerosis. J Neurol NeurosurgPsychiatry, 47, 250—255.Grant, I. (1985) The social environment and neurologicaldisease. Adv psychsom med, 13, 26—48.Green, J., Shellenberger, R. (1991) The dynamics of healthand wellness: a biopsychosocial approach. Toronto: HoltRinehart and Winston Inc.Grinker, R.R., Ham, G.C., Robbins, F.P. (1950) Somepsychodynamic factors in multiple sclerosis. Proc AssnRes Nerv Ment Dis, 28, 456-460.Guttman, L. (1945) A basis for analyzing test-retestreliability. Psychometrika. 10(4), 255—282.Haistead, W.C. (1947) Brain and Intelligence: aquantitative study of the frontal lobes. Chicago:University of Chicago Press.Harrell, L.E., Duvall, E., Folks, D.G., Duke, L.,Bartolucci, A., Conboy, T., Callaway, R., Kerns, D.(1991) The relationship of high-intensity signals onmagnetic resonance images to cognitive and psychiatricstate in Alzheimer’s Disease. Arch Neurol, 48, 1136-1140.Harrower, M.R. (1950) The results of psychometric andpersonality tests in multiple sclerosis. Proc Assn ResNerv Ment Dis, 28, 461—470.160Harrower, M.R., Kraus, J. (1951) Psychological studies onpatients with multiple sclerosis. Arch NeurolPsychiatry, 66, 44-58.Harvey, I., Ron, M., du Boulay, G., Murray, R., Lewis, S.(1990) White matter abnormalities on magnetic resonanceimaging in bipolar patients. Arch Gen Psychiatry, 47,1172—1173.Hashimoto,S.A., Paty,D.W. (1986) Multiple Sclerosis.pp. 518-589, chicago: Year Book Medical Publishers.Hays, W.L. (1988) Statistics. Orlando: Holt Rinehart andWinston Inc.Heaton, R.K., Nelson, L.M., Thompson, D.S., Burks, J.S.,Franklin, G.M. (1985) Neuropsychological findings inrelapsing-remitting and chronic-progressive multiplesclerosis. J Consult Clin Psychol, 53, 103—110.Honer, W.G., Hurwitz, T., Li, D.K.B., Palmer, M., Paty,D.W. (1987) Temporal lobe involvement in multiplesclerosis patients with psychiatric disorders. ArchNeurol, 44, 187—190.Huber, S.J., Paulson, G.W., Shuttleworth, E.C., Chakeres,D., Clapp, L.E., Pakalnis, A., Weiss, K, Rammohan, K.(1987) Magnetic resonance imaging correlates ofdementia in multiple sclerosis. Arch Neurol, 732—736.Hunt, A.L., Orrison, W.W., Yeo, R.A., Haaland, K.Y., Rhyne,R.L., Garry, P.J., Rosenberg, G.A. (1989) Clinicalsignificance of MRI white matter lesions in theelderly. Neurology, 39, 1470—1474.Inman, W.S. (1948) Can emotional conflict inducedisseminated sclerosis? Br J Med Psych, 21, 135—154.Isaac, C., Li, D.K.B., Genton, M., Jardine, C., Grochowski,E., Palmer, M., Kastrukoff, L.F., Oger, J., Paty, D.W.(1988) Multiple sclerosis: A serial study using MRI inrelapsing patients. Neurology, 38, 1511-1515.Ivnik, R. (1978) Neuropsychological stability in multiplesclerosis. J Consult Clin Psychol, 46, 913-923.Jackson, J. Hughlings. (1884) Evolution and dissolution ofthe nervous system, Brit Med J, i. Re-printed in J.Taylor, ed., (1958) Selected Writings of John HughlingsJackson. vol. II, pp. 45-75, New York: Basic Books Inc.Jam, A.K., Dubes, R.C. (1988) Algorithms for ClusteringData. New Jersey: Prentice Hall Inc.161Jambor, K.L. (1969) cognitive functioning in multiplesclerosis. Br J Psychiatry, 115, 765—775.Jelliffe, S.E. (1921) Multiple sclerosis andpsychoanalysis, Am J Med Sciences, 161, 666-675.Joffe, R.T., Lippert, G.P., Gray, T.A., Sawa, G., Horvath,Z. (1989) Mood disorders and multiple sclerosis. In K.Jensen et al, eds., Mental Disorders and CognitiveDeficits in Multiple Sclerosis. pp. 121-127, London:John Libbey.Jouvent, R., Montreuil, M., Benoit, N., Lubetzki, C.,Tournier—Lasserve, E., des Lauriers, A., Widlocher, D.,Lhermitte, F., Lyon-Caen, 0. (1989) Cognitiveimpairment, emotional disturbances and duration ofmultiple sclerosis. In K. Jensen et al, eds., MentalDisorders and cognitive Deficits in Multiple Sclerosis.pp. 139-146, London: John Libbey.Kaplan, H.I., Sadock, B.J. (1991) Synopsis of Psychiatry.Baltimore: Williams and Wilkins.Keefe, F.J., Brown, G.K., Wallston, K.A., Caldwell, D.S.(1989) coping with rheumatoid arthritis pain:catastrophizing as a maladaptive strategy. Pain, 37(1),51—56.Kerlinger, F.N. (1986) Foundations of Behavioral Research.3rd ed., Fort Worth: Holt Rinehart and Winston Inc.Kertesz, A., Black, S.E., Tokar, G., Benke, T., Carr, T.,Nicholson, L. (1988) Periventricular and subcorticalhyperintensities on magnetic resonance imaging: ‘Rims,caps, and unidentified bright objects’. Arch Neurol,45, 404—408.Kessler, R.c., Price, R.H., Wortman, C.B. (1985) Socialfactors in psychopathology: Stress, social support, andcoping processes. Ann Rev Psychol, 36, 531-572.Kiloh, 1G. (1961) Pseudo-dementia. Acta Psych Scand, 37,336—351.Kionoff, H., Clark, C., Oger, J., Paty, D., Li, D. (1991)Neuropsychological performance in patients with mildmultiple sclerosis. J NervMent Dis, 179(3) 127—131.Kolb, B., Whishaw, I. Q. (1980) Fundamentals of HumanNeuropsychology. San Francisco: W.H. Freeman and Co.Kohlberg, L. (1981) The development of modes of moralthinking and choice in the years 10 to 16, chicago:University of Chicago.162Koopmans, R.A., Li, D.K.B., Grochowski, E., Cutler, P.J.,Paty, D.W. (1989) Benign versus chronic progressivemultiple sclerosis: magnetic resonance imagingfeatures. Ann Neurol, 25, 74-81.Kubler-Ross, E. (1969) on Death and Dying. New York: TheMacmillan Co.Kurtzke, J.F. (1965) Further notes on disability evaluationin multiple sclerosis, with scale modifications.Neurology, 15, 654—661.Kurtzke, J.F. (1983) Rating neurologic impairment inmultiple sclerosis: An expanded disability status scale(EDSS) . Neurology, 33, 1444—1452.Lai, C. (1982) Hierarchical grouping analysis with optionalcontiguity constraint. Vancouver: UBC—CGroup.Langworthy, O.R. (1950) A survey of the maladjustmentproblems in multiple sclerosis and the possibilities ofpsychotherapy. Proc Assn Res Nerv Ment Dis., 28, 598-611.Larcombe, N.A., Wilson, P.H. (1984) An evaluation ofcognitive—behaviour therapy for depression in patientswith multiple sclerosis. Br J Psychiatry, 145, 366-371.Lazarus, R.S., Folkman, S. (1984) Stress, appraisal andcoping. New York: Springer Verlag.Levinson, D.J. (1978) The Seasons of a Man’s Life. NewYork: Ballantine Books.Lezak, M.D. (1983) Neuropsychological Assessment. 2nd ed.Oxford Univ Press.Lishman, W.A. (1987) Organic Psychiatry. 2nd ed., Oxford:Blackwell Scientific Publications.Litvan, I., Grafman, J., Vendrell, P., Martinez, J.M.(1988) Slowed information processing in multiplesclerosis. Arch Neurol, 45, 281—285.Logsdail, S.J., Callanan, M.M., Ron, M.A. (1988)Psychiatric morbidity in patients with clinicallyisolated lesions of the type seen in multiplesclerosis: a clinical and MRI study. Psychol Med, 18,355—364.Lyon—Caen, 0., Jouvent, R., Hauser, S., Chaunu, M—P.,Benoit, N., Widlocher, D., Lhermitte, F. (1986)Cognitive functions in recent—onset demyelinatingdiseases. Arch Neurol, 43, 1138—1141.163Mahier, N.E., Davis, R.J., Benson, D.F. (1989) Screeningmultiple sclerosis patients for cognitive impairment.In K. Jensen et al, eds., Mental Disorders andCognitive Deficits in Multiple Sclerosis. pp. 11—14,London: John Libbey.Marsh, G.G., Hirsch, S.H., Leung, G. (1982) Use and misuseof the MNPI in multiple sclerosis. Psychol Rep, 51,1127—1134.Matson, R.R., and Brooks, N.A. (1977) Adjusting to multiplesclerosis: an exploratory study. Soc Sci & Ned, 11,245—250.Matthews, C.G., Cleeland, C.S., Hopper, C.L. (1970)Neuropsychological patterns in multiple sclerosis. DisNerv Syst, 31(3) , 161—170.Matthews, W.B., Acheson, E.D., Batchelor, J.R., Weller,R.O. (1985) NcAlpine’s Multiple Sclerosis. Edinburgh:Churchill Livingstone.Mclvor, G.P., Riklan, M., Reznikoff, M. (1984) Depressionin multiple sclerosis as a function of length andseverity of illness, age, remissions, and perceivedsocial support. J Clin Psychol, 40(4) 1028—1033.McFarlane, A.C., Kalucy, R.S., Brooks, P.M. (1987)Psychological predictors of disease course inrheumatoid arthritis, J Psychosomatic Res, 31(6), 757-764.Medaer, R. (1979) Does the history of multiple sclerosis goback as far as the 14th Century? Acta neurol scand, 60,189—192.Meehi, P. (1973) Construct validity in psychological tests,in Psychodiagnosis. Minneapolis: University ofMinnesota Press.Mei-Tal, V., Meyerowitz, S., Engel, G.L. (1970) The role ofpsychological process in a somatic disorder: multiplesclerosis. Psychosomatic Med, 67—86.Messick, S. (1989) Validity. In: Educational Measurement.3rd ed., New York: Macmillan Publishing.Milligan, G.W. (1980) An examination of the effect of sixtypes of error perturbation of 15 clusteringalgorithms. Psychometrika, 45(3), 325—342.164Minden, S.L., Cray, J., Reich, P. (1989) Characteristicsand predictors of depression in multiple sclerosis. InK. Jensen et al, eds., Mental Disorders and CognitiveDeficits in Multiple Sclerosis. pp. 129-136, London:John Libbey.Minden, S.L., Schiffer, R.B. (1990) Affective disorders inmultiple sclerosis: Review and recommendations forclinical research. Arch Neurol, 47, 98-104.Minden, S.L., Moes, E.J., Orav, J., Kaplan, E., Reich, P.(1990) Memory impairment in multiple sclerosis. J ClinExp Neuropsych, 12(4), 566—586.Minden,S.L., Moes,E. (1990) A psychiatric perspective.In S.M. Rao, ed., Neurobehavioral Aspects of MultipleSclerosis. pp. 230-250, New York: Oxford UniversityPress.Moore, A.D., Stambrook, M.., Peters, L.C. (1989) Copingstrategies and adjustment after closed—head injury: acluster analytic approach. Brain Injury, 3(2), 171-175.Morris, R., Blashfield, R., Satz, P. (1981) Neuropsychologyand cluster analysis: Potentials and problems. J dinNeuropsychol, 3(1), 79—99.Myers, D.G. (1989) Psychology. 2nd ed., New York: WorthPublishers Inc.Nie, H. et al (1983) Statistical Package for the SocialSciences (SPSS). New York: McGraw Hill Co.Nockleby, D.M., Deaton, A.V. (1987) Denial versus distress:coping patterns in post head trauma patients. mt Jdin Neuropsychol, IX,4, 145—148.Nott, P.N., Fleminger, J.J. (1975) Presenile dementia: thedifficulties of early diagnosis. Acta psychiat scand,51, 210—217.Ombredane, A. (1929) Sur les troubles mentaux de lascierose en plaques. Paris: Les presses universitairesde France.Orinerod, I.E.C., du Boulay, E.P.G.H., Callanan, M.M.,Johnson, G., Logsdail, S.J., Moseley, 1.S., Rudge, P.(1984) NMR in multiple sclerosis and cerebral vasculardisease. Lancet, ii, 1334.165Ormerod, I.E.C., Bronstein, A., Rudge, P., Johnson, G.,MacManus, D., Halliday, A.M., Barratt, H., du Boulay,E.P.G.H., Kendal, B.E., Moseley, I.F., Jones, S.J.,Kriss,A., Peringer, E. (1986) Magnetic resonanceimaging in clinically isolated lesions of the brainstem. J Neurol Neurosurg Psychiatry, 49, 737-743.Ormerod, I.E.C., Miller, D.H., McDonald, W.I., du Boulay,E.P.G.H., Rudge, P., Kendall, B.E, Moseley, I.F.,Johnson, G., Tofts, P.S., Halliday, A.M., Bronstein,A.M., Scaravilli, F., Harding, A.E., Barnes, D., Zilka,K.J. (1987) The role of NMR imaging in the assessmentof multiple sclerosis and isolated neurologicallesions. Brain, 110, 1579—1616.Paty, D.W., Bergstrom, M., Palmer, M., MacFadyen, J., Li,D. (1985) A quantitative magnetic resonance image ofthe multiple sclerosis brain (abstract). Neurology, 35,supp 1, 137.Paty, D.W., Oger, J.J.F., Kastrukoff, L.F., Hashimoto, S.A.,Hooge, J.P., Eisen, A.A., Eisen, K.A., Purves, S.J.,Low, M.D., Brandejs, V., Robertson, W.D., Li, D.K.B.(1988) MRI in the diagnosis of MS: a prospective studywith comparison of clinical evaluation, evokedpotentials, oligoclonal banding, and CT. Neurology, 38,180—185.Paty, D.W., Li,D.K.B. (1988) Neuroimaging in multiplesclerosis. In Clinical Neuroimaging, pp.249-278, AlanR. Liss, Inc.Peyser, J.M., Edwards, K.R., Poser, C.M. (1980a)Psychological profiles in patients with multiplesclerosis. Arch Neurol, 37, 437—440.Peyser, J.M., Edwards, K.R., Poser, C.M., Filskov, S.B.(1980b) Cognitive function in patients with multiplesclerosis. Arch Neurol, 37, 577—579.Peyser, J.M., Becker, B. (1984) Neuropsychologic evaluationin patients with multiple sclerosis. In C.M. Poseret al, eds., The Diagnosis of Multiple Sclerosis.New York: Thieme-Stratton.Peyser, J.M., Poser, C.M. (1986) Neuropsychologicalcorrelates of multiple sclerosis. In S.B. Filskov, T.J.Boll, eds., Handbook of Clinical Neuropsychology,volume 2, New York: J. Wiley & Sons.Peyser, J.M., Rao, S.M., LaRocca, N.G., Kaplan, E., C.M.Poser, éd. (1990) Guidelines for neuropsychologicalresearch in multiple sclerosis. Arch Neurol, 47, 94—97.166Philippopolous, G.S., Wittkower, E.D., Cousineau, A. (1958)The etiological significance of emotional factors inonset and exacerbations of multiple sclerosis.Psychosomatic Med, 20(6), 458—474.Piaget, J. (1929) The child’s conception of the world.London: Routledge and Kegan Paul Ltd.Pierce, R.V. (1909) The People’s Common Sense MedicalAdviser. 72nd ed., Buffalo: World’s Dispensary PrintingOffice and Bindery.Pincus, T., Callahan, L.F., Bradley, L.A., Vaughn, W..K.,Wolfe,F. (1986) Elevated M[4P1 scores forhypochondriasis, depression and hysteria in patientswith rheumatoid arthritis reflect disease rather thanpsychological status. Arthritis Rheum, 29(12), 1456-1466.Poser, C.M., Paty, D.W., Scheinberg, L., McDonald, W.I.,Davis, F.A., Ebers, G.C., Johnson, K.P., Sibley, W.A.,Silberberg, D.H., Tourtellotte, W.W. (1983) Newdiagnostic criteria for multiple sclerosis: Guidelinesfor research protocols. Ann Neurol, 13(3), 227—231.Poser, C.M., Paty, D.W., Scheinberg, L., McDonald, W.I.,Ebers,G. (1984) The Diagnosis of Multiple Sclerosis.New York: Thieme-Stratton.Pratt, R.T.C. (1951) An investigation of the psychiatricaspects of disseminated sclerosis. J Neurol NeurosurgPsychiatry, 14, 326—335.Prigatano, G.P. (1987) Psychiatric aspects of head injury:Problems areas and suggested guidelines for research.In H.S. Levin, H.M. Eisenberg, eds., NeurobehaviouralRecovery from Head Iniury. pp. 215-231, New York:Oxford University Press.Pyke, S.W., Agnew, N.M. (1991) The Science Game: anintroduction to research in the social sciences. NewJersey: Prentice—Hall Inc.Rabins, P.V., Brooks, B.R., O’Donnell, P., Pearlson, G.D.,Moberg, P., Jubelt, B., Coyle, P., Dabs, N., Folstein,M.F. (1986) Structural brain correlates of emotionaldisorder in multiple sclerosis. Brain, 109, 585—597.Rao, S.M., Hammeke, T.A., McQuillen, M.P., Khatri, B.O.,Lloyd, D. (1984) Memory disturbance in chronicprogressive multiple sclerosis. Arch Neurol, 41, 625—631.167Rao, SM., Glatt, S., Hammeke, T.A., McQuillen, M.P.,Khatri, B.C., Rhodes, A.M.., Pollard, S. (1985) Chronicprogressive multiple sclerosis: Relationship betweencerebral ventricular size and neuropsychologicalimpairment. Arch Neurol, 42, 678—682.Rao, S.M. (1986) Neuropsychology of multiple sclerosis: Acritical review. J Clin Exp Neuropsychol, 8(5), 503-542.Rao, S.M., Leo, G.J., Haughton, V.M., St. Aubin-Faubert, P.,Bernardin, L. (1989a) Correlation of magneticresonance imaging with neuropsychological testing inmultiple sclerosis. Neurology, 39, 161—166.Rao, S.M., Bernardin, L., Leo, G.J., Ellington, L., Ryan,S.B., Burg, L.S. (1989b) Cerebral disconnection inmultiple sclerosis: Relationship to atrophy of thecorpus callosum. Arch Neurol, 46, 918—920.Rao, S.M., Mittenberg, W., Bernardin, L., Haughton, V., Leo,G.J. (1989c) Neuropsychological test findings insubjects with leukoaraiosis. Arch Neurol, 46, 40-44.Rao, S.M., St. Aubin-Faubert, P., Leo, G.J. (1989d)Information processing speed in patients with multiplesclerosis. J Clin Exp Neuropsychol, 11(4) 471-477.Rao, S.M., ed. (1990) Neurobehavioral Aspects of MultipleSclerosis. New York: Oxford University Press.Riklan, M., Levita, E., Diller, L. (1961) Psychologicstudies in neurologic diseases — a review: Parkinson’sdisease and multiple sclerosis. J Am Ger Soc, 9, 857—867.Robinson, R.G., Price, T.R. (1982) Post-stroke depressivedisorders: a follow-up study of 103 patients. Stroke,13(5), 635—640.Ron, M.A., Logsdail, S.J. (1989) Psychiatric morbidity inmultiple sclerosis: a clinical and MRI study. PsycholNed, 19, 887—895.Ross, A.T., Reitan, R.M. (1955) Intellectual and affectivefunctions in multiple sclerosis. Arch NeurolPsychiatry, 73, 663—677.Schiffer, R.B., Caine, E.D., Bamford, K.A., Levy, S.(1983) Depressive episodes in patients with multiplesclerosis. Am J Psychiatry, 140(11), 1498—1500.168Schiffer,R.B. (1987) The spectrum of depression in multiplesclerosis. Arch Neurol, 44, 596—599.Schiffer, R.B., Wineman, N.M. (1990) Antidepressantpharmacotherapy of depression associated with multiplesclerosis. Am J Psychiatry, 147, 1493—1497.Seguin, E.C. (1878) A contribution to the pathologicalanatomy of disseminated cerebro—spinal sclerosis.J Nerv Ment Dis, 5, 281-293.Shontz, F.C. (1955) MMPI responses of patients withmultiple sclerosis. J Consult Psychol, 19(1), 74.Skegg, K., Corwin, P.A., Skegg, D.C.G. (1988) How often ismultiple sclerosis mistaken for a psychiatric disorder?Psychol Ned, 18, 733—736.Stenager,E., Knudson,L., Jensen,K. (1989a) Historical noteson mental aspects of multiple sclerosis, in K. Jensenet al, eds., Mental Disorders and Cognitive Deficits inMultiple Sclerosis. pp. 1-7, London: John Libbey.Stenager, E., Knudsen, L., Jensen, K. (1989b) Multiplesclerosis: correlation of cognitive dysfunction withKurtzke disability status scale, in K. Jensen et al,eds., Mental Disorders and Cognitive Deficits inMultiple Sclerosis. pp. 27-37, London: John Libbey.Stevens, J.C., Farlow, N.R., Edwards, M.K., Yu, P-l. (1986)Magnetic resonance imaging: Clinical correlation in 64patients with multiple sclerosis. Arch Neurol, 43,1145—1148.Stewart, W.A., Hall, L.D., Berry, K., Paty, D.W. (1984)Correlation between NMR scan and brain slice data inmultiple sclerosis. Lancet, 412.Stewart, W.A., Hall, L.D., Berry, K., Churg,A., Oger,J.,Hashimoto,S.A., Paty,D.W. (1986) Magnetic resonanceimaging (MRI) in multiple sclerosis (MS): pathologicalcorrelation studies in eight cases. Neurology 36, supp1, 320.Strauss, D.H., Spitzer, R.L., Muskin, P.R. (1990)Naladaptive denial of physical illness: A proposal forDSM-IV. Am J Psychiatry, 147, 1168-1172.Strub, R.L., Black, F.W. (1985) The Mental StatusExamination in Neurology. 2nd ed., Philadelphia:F.A.Davis Co.Sugar, C., Nadell, R. (1943) Mental symptoms in multiplesclerosis. J Nerv Ment Dis, 98, 267—280.169Surridge, D. (1969) An investigation into some psychiatricaspects of multiple sclerosis. Br J Psychiatry, 115,749—764.Talbott, J.A., Hales, R.E., Yudofsky, s.c. (1988) Textbook —of Psychiatry. Washington: American Psychiatric PressInc.Taylor, S.E. (1991) Health Psychology. 2nd ed., New York:McGraw-Hill Inc.Taylor, J. (1958) Selected Writings of John HughlingsJackson. vol. II. New York: Basic Books Inc.Trimble, M.R., Grant, I. (1982) Psychiatric aspects ofmultiple sclerosis. In D.F. Benson, D. Blumer, eds.,Psychiatric aspects of Neurological Disease, vol II,pp. 279-299, New York: Grune and Stratton.Valk, J., van der Knaap, M.S. (1989) Magnetic ResonanceImaging of Myelin, Myelination and Myelin Disorders.Berlin: Springer-Verlag.Van den Burg, W., van Zoineren, A.H., Minderhoud, J.M.,Prange, A.J.A., Meijer, N.S.A. (1987) cognitiveimpairment in patients with multiple sclerosis and mildphysical disability. Arch Neurol, 44, 494—501.VanderPlate, C. (1984) Psychological aspects of multiplesclerosis and its treatment: toward a biopsychosocialperspective. Health Psychol, 3(3), 253—272.Walsh, K. (1987) Neuropsychology: a clinical approach. 2nded., New York: churchill Livingstone.Ward, J.H. (1963) Hierarchical grouping to optimize anobjective function. Amer Stat Soc J, 236-244.Wechsler, D. (1981) WAIS-R Manual: Wechsler AdultIntelligence Scale — revised. San Antonio, Texas: ThePsychological Corp., Harcourt Brace Jovanovich.Wechsler, I.S. (1922) Statistics of multiple sclerosis, AMAArch Neurol Psychiatry, 8, 59-75.Weingartner, H., Silberman, E. (1982) Models of cognitiveimpairment: Cognitive changes in depression.Psychopharm Bull, 18(2), 27—42.Weinstein, E.A. (1970) Behavioral aspects of multiplesclerosis. Modern Treatment, 8, 961-968.170Whitlock, F.A., Siskind, M.M. (1980) Depression as a majorsymptom of multiple sclerosis. J Neurol NeurosurgPsychiatry, 43, 861-65.Willoughby, E.W., Paty, D.W. (1988) Scales for ratingimpairment in multiple sclerosis: a critique.Neurology, 38, 1793—1798.Winer, B.J. (1971) Statistical Principles in ExperimentalDesign. 2nd ed., New York: McGraw Hill Book Co.Yetkin, F.Z., Haughton, V.M., Papke, R.A., Fischer, N.E.,Rao, S.M. (1991) Multiple sclerosis: Specificity of MRfor diagnosis. Radiology, 178(2), 447—451.Young, A.C., Saunders, J., Ponsford, J.R. (1976) Mentalchange as an early feature of multiple sclerosis. JNeurol Neurosurg Psychiatry, 39, 1008-1013.Zeldow, P.B., Pavlou, N. (1988) Physical and psychosocialfunctioning in multiple sclerosis: Descriptions,correlations and a tentative typology. Br J MedPsychol, 61, 185—195.171


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items