UBC Faculty Research and Publications

A startling acoustic stimulus interferes with upcoming motor preparation : evidence for a startle refractory… Maslovat, Dana; Chua, Romeo; Carlsen, Anthony N.; May, Curtis; Forgaard, Christopher J.; Franks, Ian M. Jun 30, 2015

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

Item Metadata


52383-Maslovat_D_et_al_Startling_acoustic_stimulus.pdf [ 458.94kB ]
JSON: 52383-1.0132656.json
JSON-LD: 52383-1.0132656-ld.json
RDF/XML (Pretty): 52383-1.0132656-rdf.xml
RDF/JSON: 52383-1.0132656-rdf.json
Turtle: 52383-1.0132656-turtle.txt
N-Triples: 52383-1.0132656-rdf-ntriples.txt
Original Record: 52383-1.0132656-source.json
Full Text

Full Text

Startle Refractory Period    1 Running head: Startle Refractory Period 1  2 A startling acoustic stimulus interferes with upcoming motor preparation: Evidence for a 3 startle refractory period 4  5  6 Dana Maslovat1, 2, Romeo Chua1, Anthony N. Carlsen3, Curtis May1, Christopher J. 7 Forgaard1, & Ian M. Franks1 8  9 1. School of Kinesiology, University of British Columbia 10 2. Department of Kinesiology, Langara College 11 3. School of Human Kinetics, University of Ottawa 12  13  14 Address Correspondence to: 15 Dana Maslovat 16 School of Kinesiology, University of British Columbia 17 War Memorial Gymnasium 210-6081 University Boulevard  18 Vancouver, British Columbia, CANADA, V6T 1Z1 19 Office: (604) 822-3400 Fax: (604) 822-6842 20 Email: dmaslovat@langara.bc.ca  21 19/10/2015 22  23 24 Startle Refractory Period    2 Abstract 25 When a startling acoustic stimulus (SAS) is presented in a simple reaction time 26 (RT) task, response latency is significantly shortened. The present study used a SAS in a 27 psychological refractory period (PRP) paradigm to determine if a shortened RT1 latency 28 would be propagated to RT2. Participants performed a simple RT task with an auditory 29 stimulus (S1) requiring a vocal response (R1), followed by a visual stimulus (S2) 30 requiring a key-lift response (R2). The two stimuli were separated by a variable stimulus 31 onset asynchrony (SOA), and a typical PRP effect was found. When S1 was replaced 32 with a 124 dB SAS, R1 onset was decreased by 40-50 ms; however, rather than the 33 predicted propagation of a shortened RT, significantly longer responses were found for 34 RT2 on startle trials at short SOAs. Furthermore, the 100 ms SOA condition exhibited 35 reduced peak EMG for R2 on startle trials, as compared to non-startle trials. These results 36 are attributed to the startling stimulus temporarily interfering with cognitive processing, 37 delaying and altering the execution of the second response. In addition to this “startle 38 refractory period,” results also indicated that RT1 latencies were significantly lengthened 39 for trials that immediately followed a startle trial, providing evidence for longer-term 40 effects of the startling stimulus.  41  42 Keywords: psychological refractory period, dual-task performance, response preparation, 43 startle reflex 44  45  46 47 Startle Refractory Period    3 1. Introduction 48 A common technique used over the past century to examine people’s ability to 49 perform multiple activities concurrently is the psychological refractory period paradigm 50 (Telford, 1931), in which participants are required to identify and respond to two stimuli 51 (S1 and S2) which are separated in time. Typically, as the time interval between the two 52 stimuli (stimulus onset asynchrony; SOA) shortens, the reaction time (RT) to respond to 53 the first stimulus (RT1) is unaffected, while the response latency to the second stimulus 54 (RT2) is increased. The delay in RT2 is known as the psychological refractory period 55 (PRP) and is thought to be indicative of the cost associated with processing two stimulus-56 response streams simultaneously (see Lien & Proctor, 2002; Pashler, 1994; 1998 for 57 reviews). 58 Explanations offered for a delayed RT2 in PRP tasks can typically be divided into 59 capacity sharing or “bottleneck” models (Pashler, 1994). Capacity theories assume that 60 processing resources are shared among tasks and thus when multiple tasks are performed 61 there is less resource available for each task, leading to impaired performance 62 (Kahneman, 1973). Conversely, bottleneck theories posit that certain processing stages 63 cannot be performed in parallel and thus processing multiple stimuli reaches a rate-64 limiting stage at some point whereby only one item can be processed at a time. Although 65 the location of the bottleneck is still debated, considerable evidence exists suggesting that 66 stimulus perception can occur in parallel and therefore is unlikely to contribute to the 67 bottleneck (Pashler, 1994). While some research has provided support for a response 68 selection bottleneck (e.g., Karlin & Kestenbaum, 1968; Smith, 1969), a PRP effect also 69 occurs in a simple RT paradigm where response selection is minimal, indicating the 70 Startle Refractory Period    4 bottleneck may involve the response production stage (Bratzke, Rolke, & Ulrich, 2009; 71 Maslovat, et al., 2013). It is also possible that a bottleneck occurs at multiple stages or 72 that a central bottleneck affects both response selection and movement production (De 73 Jong, 1993; Pashler, 1994). 74 In order to examine the PRP effect and which stage of processing is affected, the 75 bottleneck theory offers a number of testable predictions. One such prediction is that any 76 modification to task 1 that changes the central processing time required (up to or 77 including the bottleneck stage), should have an equal effect on both RT1 and RT2 78 (Pashler, 1994). That is, at short SOAs, any RT change of task 1 should be propagated to 79 task 2 (see Figure 3, middle panel), whereas propagation effects would not be predicted at 80 long SOAs as there is no overlap in processing (Miller & Reynolds, 2003). Propagation 81 effects have been confirmed by manipulating response selection variables such as number 82 of response alternatives (Karlin & Kestenbaum, 1968; Smith, 1969), as well as response 83 production variables such as sequence length (Bratzke, et al., 2008) or movement 84 amplitude (Bratzke, et al., 2009; Ulrich, et al., 2006). In these experiments, increasing the 85 time required to process task 1 resulted in similar magnitude increases for both RT1 and 86 RT2 at short SOAs, consistent with the predictions of the bottleneck theory. Additionally, 87 other research has reduced the response latency of RT1 through increased temporal 88 predictability (Bausenhart, Rolke, Hackley, & Ulrich, 2006) or practice (Ruthruff, 89 Johnston, Van Selst, Whitsell, & Remington, 2003), resulting in a similar decrease in 90 RT2 at short SOAs. 91 The purpose of the current study was to examine response propagation effects in a 92 PRP paradigm by reducing task 1 latency through the use of a startling acoustic stimulus 93 Startle Refractory Period    5 (SAS). When a SAS is presented in a simple RT task, RT is significantly shortened as the 94 SAS acts as an involuntary trigger of the prepared response, bypassing response selection 95 processes and shortening stimulus detection and response initiation stages (see Carlsen, 96 Maslovat, & Franks, 2012; Valls-Solé, Kumru, & Kofler, 2008 for reviews). Specifically, 97 it is thought that the SAS activates subcortical brain structures via connections between 98 the cochlear nucleus and reticular formation, leading to both a reflexive startle response 99 as well as involuntary activation leading to the initiation of a prepared response (provided 100 a sufficient level of advance preparation of the movement; see Carlsen, et al., 2012 for 101 more details). As the pathways and processes associated with the startle-mediated release 102 of a response are faster than voluntary response initiation, responses to the SAS are 103 significantly shortened as compared to non-startle trials (e.g., muscle activation onset <80 104 ms; Valls-Solé, Rothwell, Goulart, Cossu, & Munoz, 1999).  105 In the current study, participants performed two simple RT tasks in a PRP 106 paradigm, in which they were required to respond to an auditory stimulus (S1) with a 107 vocal response (R1), which was followed by a visual stimulus (S2) requiring a key-lift 108 movement (R2). On selected trials, S1 was replaced with a SAS, with the expectation that 109 this would shorten RT1 latency in the range of 40-60 ms, as has been previously shown 110 for a vocal response (Stevenson, et al., 2014). Of primary interest was whether the RT 111 “savings” associated with startle trials would propagate to RT2 for short SOAs, as 112 predicted by the central bottleneck model. As both responses were known in advance, 113 any propagation effects would be attributed to a shortened response execution stage of 114 R1, leading to a similar reduction in the latency of R2. Although this logic is similar to 115 previous work examining propagation effects, the use of a SAS provides unique benefits, 116 Startle Refractory Period    6 as the SAS is considered to act via a separate and involuntary response initiation 117 pathway, thus bypassing any response initiation bottleneck (Bratzke, et al., 2009; De 118 Jong, 1993). Indeed, a SAS has been successfully used in a dual-task paradigm to assess 119 the attentional demands of a continuous task (Begeman, Kumru, Leenders, & Valls-Sole, 120 2007), as well as in a PRP paradigm as a probe to determine the preparation level of the 121 second response (Maslovat, et al., 2013). 122 2. Methods 123 2.1 Participants 124 Data were collected from 17 right-handed volunteers with no sensory or motor 125 dysfunctions. However, five participants were excluded due to a lack of activation in the 126 sternocleidomastoid (SCM) muscle within 120 ms following a SAS (a reliable indicator 127 of a startle response; see Carlsen, Maslovat, Lam, Chua, & Franks, 2011 for inclusion 128 criteria) on all four startle trials in the single-task vocal RT block (see Section 2.2 129 Experimental Design). Thus, data are presented from twelve participants (7 male, 5 130 female; M = 24.8 yrs, SD = 6.1 yrs). All participants signed an informed consent form 131 and were naïve to the hypothesis under investigation. This study was approved by the 132 University of British Columbia ethics committee and was conducted in accordance with 133 the ethical guidelines set forth by the Declaration of Helsinki.  134 2.2 Apparatus, Task, and Experimental Design 135 Participants sat in a height-adjustable chair in front of a table with a 22-inch 136 computer monitor (Acer X233W, 1152 x 864 pixels, 75 Hz refresh) placed on it. 137 Participants placed the right hand on a telegraph key (E.F. Johnson Speed-X, Model 114-138 300) located on the table that required 2 N of force to close (i.e., simply resting the hand 139 Startle Refractory Period    7 on the switch was sufficient to close it). A microphone (Sennheiser, MKH 416-P48) was 140 placed in front of the participant, below the monitor to capture vocal responses.  141 To determine baseline performance, participants began by performing 20 trials of 142 each of the two required responses in a single-task situation. All trials began with the 143 word “Ready!” presented on the computer screen, followed by a variable foreperiod of 144 2500-3500 ms. For the first block of trials, participants were instructed to respond to an 145 auditory stimulus by vocalizing the word “TAT” as quickly as possible. The auditory 146 stimulus consisted of a non-startling tone on 16 trials (82 +/-2 dB, 40 ms, 1000 Hz) and a 147 startling tone on 4 trials (124 +/-2 dB, 40 ms, 1000 Hz, <1 ms rise time). Startle trials 148 were interspersed pseudorandomly such that the first trial was never a startle trial and 149 there were never two consecutive startle trials. Acoustic signals were generated by a 150 customized computer program and were amplified and presented via a loudspeaker 151 placed behind the head of the participant. Acoustic stimulus intensity was measured at a 152 distance of 30 cm from the loudspeaker (approximately the distance to the ears of the 153 participant) using a sound level meter (Cirrus Research model CR:252B; “A”-weighted 154 decibel scale, impulse response mode). In the second block of trials, participants were 155 instructed to respond to the presentation of a green circle (10 cm diameter) in the middle 156 of the computer screen by lifting their right hand off the telegraph key as quickly as 157 possible. During the single-task testing blocks, RT was presented on the screen for five 158 seconds following each trial with a monetary reward of CDN $0.05 per trial for RTs 159 below 250 ms. 160 Following the single-task trials, participants were informed that they would be 161 performing both the vocal response and key-lift in a dual-task situation, and that they 162 Startle Refractory Period    8 should give equal priority to performing each task as quickly as possible. The auditory 163 stimulus (S1) was always presented first and required a vocal response of “TAT” (R1), 164 followed by the visual stimulus (S2) requiring a right hand key-lift response (R2). A 165 practice block of 20 trials was conducted, with SOAs of 100 ms (10 trials), 200 ms (4 166 trials), 500 ms (2 trials), 1000 ms (2 trials), and 1500 ms (2 trials) randomly presented. A 167 high proportion of short SOA trials were used, as propagation effects are only expected 168 for these conditions. Following the practice block, participants performed 5 blocks of 25 169 test trials whereby 20 trials involved the same distribution of SOAs as the practice trials, 170 but one additional trial was presented at each SOA where the 124 dB SAS was presented 171 in place of the normal 82 dB auditory stimulus (S1) (i.e., 5 startle trials per test block, 25 172 startle trials total). Startle trials were interspersed pseudorandomly within each block in a 173 similar manner to the single-task testing condition. During the dual-task testing blocks, 174 RT for each task was presented simultaneously on the screen for seven seconds following 175 each trial with a monetary bonus of CDN $0.05 per task (i.e., up to $0.10 per trial) for 176 fast RTs (<250 ms for RT1, <300 ms for RT2). Participants were instructed to try and 177 maximize their reward bonus by minimizing total RT and thus receiving the reward 178 bonus for both responses. Participants were allowed a rest period of approximately one 179 minute in between blocks and the testing session lasted approximately one hour. 180 2.3 Recording Equipment 181 Surface EMG data were collected from the muscle bellies of the right extensor 182 carpi radialis longus (ECR - agonist), and right and left sternocleidomastoid (SCM – used 183 as a startle indicator only) using preamplified surface electrodes connected via shielded 184 cabling to an external amplifier system (Delsys Model DS-80). Recording sites were 185 Startle Refractory Period    9 prepared and cleansed in order to decrease electrical impedance. The electrodes were 186 oriented parallel to the muscle fibers, and then attached using double sided adhesive 187 strips. A grounding electrode was placed on the left ulnar styloid process. EMG onsets 188 were defined as the first point where the rectified and filtered (25 Hz low pass elliptical 189 filter) EMG activity first reached a sustained value of two standard deviations above 190 baseline levels (mean EMG activity 100 ms prior to S1), with EMG offsets determined in 191 a similar manner. EMG onset and offset points were determined using a custom 192 LabVIEW® (National Instruments Inc.) program and then visually confirmed and 193 manually adjusted (if necessary) to compensate for any errors due to the strictness of the 194 algorithm.  195 Displacement RT of key lift-off was monitored using the contact switch of the 196 telegraph key, while vocal responses were collected using the microphone placed in front 197 of the participant. Voice onset and offset was determined in an identical manner to EMG, 198 whereas displacement onset for the key-lift task was determined by the time at which 199 switch contact was broken. A customized LabView® computer program controlled 200 stimulus and feedback presentation, and initiated data collection (National Instruments, 201 PC-MIO-16E-1) at a rate of 1 kHz for 3 s, starting 500 ms prior to the presentation of the 202 S1 “go” signal.  203 2.4 Data Reduction 204 The first block of dual-task trials was not analyzed as this block was considered 205 practice and did not include a SAS. Before analyzing the results of the experimental 206 blocks (1980 total trials across participants), we discarded 46 trials (2.3 %) in which an 207 error occurred (most often due to a telegraph key not being fully depressed at the start of 208 Startle Refractory Period    10 the trial), 14 trials (0.8 %) in which a response occurred prior to the stimulus (i.e., 209 anticipation), 17 trials (1.1%) in which a slow (>500 ms) vocal response (R1) occurred , 210 and 16 trials in which the participant did not show any SCM activation within the first 211 120 ms for a startle trial (i.e., lack of startle indicator). Of the remaining 1887 trials, we 212 discarded an additional 93 trials (4.9 %) in which the two responses occurred less than 213 100 ms apart, as these trials may represent a “grouped” response which may introduce 214 unwanted effects (see Miller & Ulrich, 2008; Ulrich & Miller, 2008 for more details). 215 Overall, our analysis included 1794 of the 1980 total trials (90.6 %). 216 2.5 Dependent Measures & Analyses 217 Primary dependent measures included voice onset (RT1) and key-lift 218 displacement onset (RT2). To confirm that processing time for R1 (vocal response) was 219 not different between the single-task condition and all SOA conditions in the dual-task 220 paradigm, we analyzed RT1 via a 2 Stimulus (non-startle, startle) x 6 Condition (single-221 task, 100 SOA, 200 SOA, 500 SOA, 1000 SOA, 1500 SOA) repeated measures analysis 222 of variance (ANOVA). To confirm a typical PRP effect for the key-lift task (R2), we 223 examined RT2 for non-startle trials using a one-way, 6 factor (Condition: single-task, 100 224 SOA, 200 SOA, 500 SOA, 1000 SOA, 1500 SOA), repeated measures ANOVA. To 225 determine the effects of the SOA and startling stimulus on performance of the key-lift 226 task (R2), RT2 was analyzed using a 2 Stimulus (non-startle, startle) x 5 SOA (100 SOA, 227 200 SOA, 500 SOA, 1000 SOA, 1500 SOA) repeated-measures ANOVA.  228 We were also interested in whether the performance characteristics of the vocal 229 and key-press response were affected by either the intensity of S1 or SOA condition. 230 Thus, we measured the vocal response duration as well as ECR (agonist) duration and 231 Startle Refractory Period    11 peak amplitude (defined as maximal rectified EMG amplitude between onset and offset) 232 for the key-lift task. Voice duration was analyzed via a 2 Stimulus (non-startle, startle) x 233 6 Condition (single-task, 100 SOA, 200 SOA, 500 SOA, 1000 SOA, 1500 SOA) repeated 234 measures ANOVA, whereas ECR duration and peak amplitude were analyzed using a 2 235 Stimulus (non-startle, startle) x 5 SOA (100 SOA, 200 SOA, 500 SOA, 1000 SOA, 1500 236 SOA) repeated-measures ANOVA. 237 Greenhouse-Geisser corrected degrees of freedom were used to adjust for 238 violations of sphericity if necessary. Uncorrected degrees of freedom are reported, with 239 the corrected p values. Partial eta squared (ηp2) values are reported as a measure of effect 240 size. The alpha level for the entire experiment was set at .05, and where appropriate, 241 significant results were examined via Tukey’s honestly significant difference (HSD) test 242 to determine the locus of the differences. 243 3. Results  244 3.1 Response Latencies 245 As expected, analysis of vocal responses showed that RT1 latencies were 246 significantly shorter on startle trials (M = 172 ms, 95% CI [153.5, 190.1]) compared to 247 non-startle trials (M = 216 ms, 95% CI [193.3, 238.2]), as confirmed by a main effect of 248 stimulus, F(1, 11) = 136.56, p < .001, ηp2 = .93 (Figure 1A). Analysis of RT1 also yielded 249 a significant main effect of condition, F(5, 55) = 7.75, p =.004, ηp2 = .41 which post-hoc 250 testing confirmed was due to a significantly longer RT1 when performed as a single-task 251 compared to all conditions of the dual-task paradigm, which were not significantly 252 different to each other. This effect has been shown previously and has been attributed to 253 practice effects when the single-task paradigm is performed prior to the dual-task trials 254 Startle Refractory Period    12 (Maslovat, et al., 2013). To further confirm this main effect of condition was the result of 255 practice effects, we performed an additional post-hoc analysis of RT1 (collapsed across 256 condition) using a 2 Stimulus (non-startle, startle) x 6 Block (Single-Task, Block 1, Block 257 2, Block 3, Block 4, Block 5) repeated-measures ANOVA. This analysis produced both a 258 main effect of stimulus, F(1, 11) = 121.92, p < .001, ηp2 = .92 and a main effect of block, 259 F(5, 55) = 12.29, p < .001, ηp2 = .53, in which RT1 significantly decreased as the 260 experiment progressed in a linear manner, F(1, 11) = 19.37, p = .001, ηp2 = .64 (Figure 261 1B). Although a practice effect was present for RT1, the lack of difference in vocal 262 response latency between SOAs during the dual-task task indicates that the first response 263 was processed in a similar manner throughout the dual-task portion of the experiment.  264 (INSERT FIGURE 1 ABOUT HERE) 265 Analysis of the key-lift task (RT2) on non-startle trials showed a main effect of 266 condition, F(5, 55) = 120.31, p < .001, ηp2 = .92. This represents a typical PRP effect in 267 which RT2 latency significantly decreased with increasing SOA, reaching single-task 268 key-lift latencies at long SOAs (Figure 2). Post-hoc tests indicated that RT2 was 269 significantly longer at SOAs of 100 ms (M = 343 ms, 95% CI [316.5, 370.2]), 200 ms (M 270 = 283 ms, 95% CI [260.7, 306.0]), and 500 ms (M = 244 ms, 95% CI [225.1, 263.0]), as 271 compared to the single task RT2 (M = 196 ms, 95% CI [182.4, 209.9]; shown as a solid 272 black line in Figure 2). 273 (INSERT FIGURE 2 ABOUT HERE) 274 Our primary research question was whether the RT1 “savings” during startle trials 275 would be inherited by RT2, as would be predicted by the central bottleneck theory. 276 However, in contrast to our predictions, startle trials resulted in longer RT2 values at 277 Startle Refractory Period    13 short SOAs (Figure 2). Analysis of RT2 confirmed both a main effect of stimulus, F(1, 278 11) = 14.54, p = .003, ηp2 = .57, and SOA, F(4, 44) = 80.03, p < .001, ηp2 = .88, which 279 were superseded by a significant Stimulus x SOA interaction, F(4, 44) = 3.98, p = .024, 280 ηp2 = .27. Post hoc analysis of this interaction revealed that startle resulted in significantly 281 longer RT2 values compared to non-startle trials at short SOAs of 100 ms (startle M = 282 397 ms, 95% CI [346.0, 447.0], non-startle M = 343ms, 95% CI [316.5, 370.2]) and 200 283 ms (startle M = 319 ms, 95% CI [276.3, 360.8], non-startle M = 283ms, 95% CI [260.7, 284 306.0]).  285 Note that as opposed to the shortened RT1 latencies in startle trials being 286 propagated to RT2, RT2 latencies were in fact delayed on startle trials at short SOAs (see 287 Figure 3 for a schematic). Thus, to determine the effects of the SAS on RT2, it is 288 necessary to add the RT1 savings to the RT2 delay (Figure 4). These additive effects at 289 short SOAs can be considered a “startle refractory period” in which using a SAS to 290 trigger task 1 at an earlier latency results in a delay in initiating the second response. The 291 startle refractory period appears to be short in duration as no significant RT2 delay was 292 observed at longer SOAs (500 ms or greater). Although there are still RT1 savings 293 associated with long SOAs, these savings would not be predicted to be propagated to 294 RT2 due to the first response having passed through the central bottleneck.  295 (INSERT FIGURE 3 & 4 ABOUT HERE) 296 Contrary to our prediction, reducing the latency of the first response via 297 presentation of a SAS resulted in a delayed second response, which we attributed to a 298 startle refractory period. Although these effects had vanished by the 500 ms SOA, we 299 were interested in whether eliciting a startle reflex had a more lasting effect, which would 300 Startle Refractory Period    14 be demonstrated by a change in performance on the subsequent trial. To examine this 301 possibility we performed a post-hoc analysis of RT1 latency, irrespective of SOA 302 condition, using a paired sample t-test comparing the non-startle trial prior to and 303 following each startle trial in both the single-task and dual-task conditions. This ensured 304 we compared trials at a similar time in the experiment, although trials were omitted if a 305 startle trial was the last trial of a block (as there was no comparable post-startle trial), or 306 if the non-startle trial prior to a startle trial happened to also follow a startle trial (as 307 startle trials could be two trials apart). This analysis showed that post-startle trials were 308 performed with significantly longer latencies, as compared to pre-startle trials in both the 309 single-task condition, t(11) = -2.22, p = 0.048 (pre-startle M = 228 ms, post-startle M = 310 259 ms), and dual-task condition, t(11) = -2.64, p = 0.023 (pre-startle M = 209 ms, post-311 startle M = 222 ms). 312 3.2 Response Characteristics 313 Analysis of the voice duration (R1) showed that startle trials resulted in a 314 significantly longer vocal response (M = 171 ms, 95% CI [142.5, 198.6]) compared to 315 non-startle trials (M = 156 ms, 95% CI [133.6, 177.9]), as confirmed by a main effect of 316 stimulus, F(1, 11) = 7.73, p = .018, ηp2 = .41. No effects were found for condition, F(5, 317 55) = 3.50, p =.061, ηp2 = .24, or Stimulus x Condition interaction, F(5, 55) = 0.60, p 318 =.561, ηp2 = .05. Although the main effect of condition approached significance (p = 319 .061), examination of mean values indicated that this trend was primarily due to a longer 320 duration on single task trials (M = 177 ms) as compared to all other SOA conditions (100 321 ms SOA, M = 159 ms; 200 ms SOA, M = 158 ms; 500 ms SOA, M = 163 ms; 1000 ms 322 SOA, M = 162 ms; 1500 ms SOA, M = 160 ms). Consistent with the results of the RT1 323 Startle Refractory Period    15 analysis, the lack of difference in voice duration confirms that the first response was 324 produced in a similar manner during the dual-task testing conditions. 325 Analysis of the duration of the agonist EMG (R2) showed no effects of stimulus, 326 F(1, 11) = 0.69, p = .424, ηp2 = .06, SOA, F(4, 44) = 2.86, p =.098, ηp2 = .21, or Stimulus 327 x SOA interaction, F(4, 44) = 1.01, p =.345, ηp2 = .09. However, while analysis of peak 328 agonist EMG produced no main effects of stimulus, F(1, 11) = 0.19, p = .674, ηp2 = .02, 329 or SOA, F(4, 44) = 2.43, p =.125, ηp2 = .18, there was a significant Stimulus x SOA 330 interaction, F(4, 44) = 6.17, p =.002, ηp2 = .36. Post hoc analysis of this interaction 331 confirmed the only statistically different value was a significantly lowered peak agonist 332 EMG on startle trials for the 100 ms SOA (M = 0.851 mV, 95% CI [0.466, 1.236]) 333 compared to non-startle trials (M = 1.013 mV, 95% CI [0.628, 1.398]).  334 3.3 Other Considerations 335 One possible confound in this experiment is that the reflexive response to a SAS 336 typically includes a blink reflex, resulting from activation in the orbicularis oculi (OOc) 337 muscle at a latency of 35-40 ms following the SAS, with a duration of 30-150 ms 338 (Blumenthal, et al., 2005; Brown, et al., 1991). This reflexive response to the SAS may 339 have resulted in participants’ eyes being closed when the visual stimulus (S2) was 340 presented at short SOAs. To examine this possibility, we recorded EMG activity from the 341 left OOc for one participant and recorded their responses using a video camera (Casio 342 EX-F1 Exilim Digital Camera, recorded at 30 fps, image size of 512 x 384 Pixels). This 343 participant showed robust OOc activation during all startle trials with an average onset 344 latency of 50 ms and offset latency of 77ms; however, video recording showed the 345 participant’s eyes closed from 66-165 ms (± 33ms due to camera speed limitations) 346 Startle Refractory Period    16 following the SAS. Thus, for the 100 ms SOA condition, it is likely that the participant’s 347 eyes were closed when the visual stimulus was presented, which may partially explain the 348 RT2 delay. However, the auditory blink reflex was completed prior to the visual stimulus 349 in the 200 ms SOA condition and thus the RT2 delay at longer SOAs was not 350 contaminated by the reflexive activation in the OOc.  351 4. Discussion 352 The purpose of the current study was to examine RT propagation effects through 353 the use of a SAS in a PRP paradigm. On non-startle trials, participants performed the 354 vocal response at a similar latency (Figure 1A) and with a consistent duration for all 355 SOAs, confirming the first response was processed in a similar manner throughout the 356 dual-task portion of the experiment. Additionally, non-startle trials showed a typical PRP 357 effect in which shorter SOAs resulted in longer RT2 latencies, while longer SOAs 358 resulted in latencies similar to the single-task condition (Figure 2). By replacing S1 with 359 a startling stimulus, we were able to trigger the prepared vocal response and reduce RT1 360 by an average of approximately 45 ms (Figure 1A). Of primary interest was whether the 361 reduction in RT1 on startle trials would propagate to RT2, as predicted by the central 362 bottleneck model. In contrast to our prediction, startle trials produced significantly longer 363 RT2 values for the 100 ms and 200 ms SOA (Figure 2). Thus, rather than propagation 364 effects, it appears that a SAS produces a “startle refractory period” that results in a delay 365 in the preparation and/or execution of upcoming responses (Figure 3). Further evidence 366 for a transient startle refractory period is provided by significantly reduced peak agonist 367 EMG activation on startle trials for the second response at the 100 ms SOA. Thus, at 368 Startle Refractory Period    17 short SOAs, the startling stimulus not only delayed the key-lift response but also reduced 369 the amount of peak muscle activation produced by the participant. 370 The length of the startle refractory period can be estimated at short SOAs by 371 considering both the RT1 savings from the early triggering of the first response and the 372 observed RT2 delay (Figure 4). While the confound of the auditory blink reflex does not 373 allow us to accurately measure the latency of RT2 at the 100 ms SOA, data from the 200 374 ms SOA condition can provide an approximation of the startle refractory period. Even 375 with the RT1 savings of 40 ms, RT2 was delayed by an additional 35 ms, meaning that 376 the second response occurred 75 ms later than would be expected without interference 377 and with propagation effects. Note that this startle refractory period appears to be 378 independent to the psychological refractory period as no differences were found between 379 startle and non-startle trials at the 500 ms SOA, yet there was still a delay in RT2, relative 380 to single task control values (i.e. PRP effect).  381 One explanation for the short-term performance decrements may relate to motor 382 cortex suppression as a number of studies have shown that a startle-evoked activation of 383 reticulo-cortical projections can transiently (~50 ms) inhibit the motor cortex 384 (Furubayashi, et al., 2000; Kuhn, Sharott, Trottenberg, Kupsch, & Brown, 2004). 385 Similarly, it has been shown that the use of a SAS during a choice RT task can cause 386 cognitive interference and give rise to more movement production errors (Carlsen, Chua, 387 Inglis, Sanderson, & Franks, 2004). For the current study, neural activation models 388 (Hanes & Schall, 1996; see also Carlsen et al., 2012; Maslovat, Hodges, Chua, & Franks, 389 2011) predict that the amount of time required to prepare and initiate a movement is 390 dependent upon the activation level of the cortex. If the SAS causes temporary inhibition 391 Startle Refractory Period    18 of the motor cortex, it would be predicted that response latency of task 2 in a PRP 392 paradigm would also be transiently delayed at short SOAs, consistent with the reported 393 results.  394 In addition to the short-term effect of the SAS on RT2, there also appeared to be a 395 longer-term effect on reduced motor preparation as RT1 latencies were significantly 396 lengthened for trials that immediately followed a startle trial. This effect was present in 397 both single-task and dual-task conditions, suggesting that this result was not related to the 398 preparation of multiple responses but rather an effect of the startling stimulus on 399 subsequent performance. These results are in line with early studies involving the effects 400 of a startling stimulus on task performance, as researchers were concerned about possible 401 adverse effects of sonic booms on pilots. Although RTs were often facilitated by the 402 SAS, transient performance decrements were found for pursuit tracking (Thackray & 403 Touchstone, 1970; Thackray, Touchstone, & Jones, 1972) and cognitive tasks such as 404 mental arithmetic (Vlasak, 1969), which lasted as long as 20-30 seconds. Whereas the 405 aforementioned startle refractory period may involve short-term inhibition of the motor 406 cortex, the longer-term performance decrements may relate to the excitation in the 407 sympathetic nervous system caused by the acoustic startle reflex (Eder, Elam, & Wallin, 408 2009), which likely requires a longer time frame to return to pre-startle levels.  409 Although we believe the results of the current study provide strong evidence that 410 the presentation of a startling stimulus interferes with motor preparation at both a short 411 (~75 ms) and long (10-15 s) time frame, we did not directly measure motor cortex or 412 sympathetic nervous system activation. Thus, it is worthwhile to consider other 413 possibilities for the reported results. One such possibility is that detection of S2 was 414 Startle Refractory Period    19 affected by a phenomenon known as “attentional blink” (Raymond, Shapiro, & Arnell, 415 1992), in which the second of two target visual stimuli is less likely to be detected when 416 it appears in close temporal proximity to the first (see Dux & Marois, 2009 for a review). 417 More recent work has shown a similar effect with a cross-modal paradigm in which the 418 first stimulus is auditory followed by a visual second stimulus (similar to the current 419 methods), and attributed the attentional blink to a similar cortical bottleneck as implicated 420 in the PRP phenomenon (Marti, Sigman, & Dehaene, 2012).  421 While we cannot definitively rule out any effects of attentional blink in the 422 current study, a number of findings suggest that this is not a sufficient explanation for our 423 reported results. First, attentional blink paradigms usually present rapid multiple visual 424 stimuli which are flashed briefly on the screen, with the second target stimulus occurring 425 at some point in the sequence following the initial target stimulus. Conversely, the 426 current study employed a single visual stimulus that remained on the screen from initial 427 presentation until the end of the trial, requiring much less stimulus recognition processing 428 which may be responsible for the cortical bottleneck. Second, one peculiarity of the 429 attentional blink effect is that exhibits what is known as “lag-1 sparing,” meaning that if 430 the second target stimulus is presented immediately following the first target stimulus 431 (rather than later in the sequence), detection is not negatively affected (Hommel & 432 Akyurek, 2005). In the current study, the stimulus following S1 was always the visual 433 “go” signal, which would thus be unlikely to be affected by the attentional blink. Third, 434 any effects of attentional blink would be present on all trials, yet our results show clear 435 effects of the SAS presentation on RT2 latency and peak EMG at the short SOA 436 condition, as well as delayed RT in the trial following a startle. Thus we believe the 437 Startle Refractory Period    20 reported results are more likely to be attributed to effects of the startling stimulus, rather 438 than other confounding factors such as the attentional blink. 439 In summary, by implementing a startling acoustic stimulus in a psychological 440 refractory period paradigm, we have provided novel evidence that a SAS interferes with 441 motor preparation of subsequent actions. This interference results in reduced preparation 442 in the short-term (~75 ms following the SAS), which we attribute to cortical suppression 443 and in the long-term (10-15 s following the SAS), which we attribute to recovery from 444 excitation of the sympathetic nervous system. 445 446 Startle Refractory Period    21 Acknowledgements 447 Acknowledgements for this study go to separate Natural Sciences and 448 Engineering Research Council of Canada grants awarded to Ian M. Franks (RGPIN-449 2014-05172) and Romeo Chua (RGPIN-2014-06051).  450 451 Startle Refractory Period    22 References  452 Bausenhart, K. M., Rolke, B., Hackley, S. A., & Ulrich, R. (2006). The locus of temporal 453 preparation effects: evidence from the psychological refractory period paradigm. 454 Psychonomic Bulletin and Review, 13, 536-542. 455 Begeman, M., Kumru, H., Leenders, K., & Valls-Sole, J. (2007). Unilateral reaction time 456 task is delayed during contralateral movements. Experimental Brain Research, 457 181, 469-475. 458 Blumenthal, T. D., Cuthbert, B. N., Filion, D. L., Hackley, S., Lipp, O. V., & Van Boxtel, 459 A. (2005). Committee report: Guidelines for human startle eyeblink 460 electromyographic studies. Psychophysiology, 42, 1-15. 461 Bratzke, D., Rolke, B., & Ulrich, R. (2009). The source of execution-related dual-task 462 interference: Motor bottleneck or response monitoring? Journal of Experimental 463 Psychology: Human Perception and Performance, 35, 1413-1426. 464 Bratzke, D., Ulrich, R., Rolke, B., Schroter, H., Jentzsch, I., & Leuthold, H. (2008). 465 Motor limitation in dual-task processing with different effectors. Quarterly 466 Journal of Experimental Psychology, 61, 1385-1399. 467 Brown, P., Rothwell, J. C., Thompson, P. D., Britton, T. C., Day, B. L., & Marsden, C. 468 D. (1991). New observations on the normal auditory startle reflex in man. Brain, 469 114, 1891-1902. 470 Startle Refractory Period    23 Carlsen, A. N., Chua, R., Inglis, J. T., Sanderson, D. J., & Franks, I. M. (2004). Can 471 prepared responses be stored subcortically? Experimental Brain Research, 159, 472 301-309. 473 Carlsen, A. N., Maslovat, D., & Franks, I. M. (2012). Preparation for voluntary 474 movement in healthy and clincial populations: Evidence from startle. Clinical 475 Neurophysiology, 123, 21-33. 476 Carlsen, A. N., Maslovat, D., Lam, M. Y., Chua, R., & Franks, I. M. (2011). 477 Considerations for the use of a startling acoustic stimulus in studies of motor 478 preparation in humans. Neuroscience and Biobehavioral Reviews, 35, 366-376. 479 De Jong, R. (1993). Multiple bottlenecks in overlapping task performance. Journal of 480 Experimental Psychology: Human Perception and Performance, 19, 965-980. 481 Dux, P. E., & Marois, R. (2009). The attentional blink: a review of data and theory. 482 Attention, Perception and Psychophysics, 71, 1683-1700. 483 Eder, D. N., Elam, M., & Wallin, B. G. (2009). Sympathetic nerve and cardiovascular 484 responses to auditory startle and prepulse inhibition. International Journal of 485 Psychophysiology, 71, 149-155. 486 Furubayashi, T., Ugawa, Y., Terao, Y., Hanajima, R., Sakai, K., Machii, K., Mochizuki, 487 H., Shiio, Y., Uesugi, H., Enomoto, H., & Kanazawa, I. (2000). The human hand 488 motor area is transiently suppressed by an unexpected auditory stimulus. Clinical 489 Neurophysiology, 111, 178-183. 490 Startle Refractory Period    24 Hanes, D. P., & Schall, J. D. (1996). Neural control of voluntary movement initiation. 491 Science, 274, 427-430. 492 Hommel, B., & Akyurek, E. G. (2005). Lag-1 sparing in the attentional blink: benefits 493 and costs of integrating two events into a single episode. Quarterly Journal of 494 Experimental Psychology. A, Human Experimental Psychology, 58, 1415-1433. 495 Kahneman, D. (1973). Attention and effort. Englewood Cliffs, NJ: Prentice-Hall. 496 Karlin, L., & Kestenbaum, R. (1968). The effects of number of alternatives on the 497 psychological refractory period. Quarterly Journal of Experimental Psychology, 498 20, 167-178. 499 Kuhn, A. A., Sharott, A., Trottenberg, T., Kupsch, A., & Brown, P. (2004). Motor cortex 500 inhibition induced by acoustic stimulation. Experimental Brain Research, 158, 501 120-124. 502 Lien, M. C., & Proctor, R. W. (2002). Stimulus-response compatibility and psychological 503 refractory period effects: implications for response selection. Psychonomic 504 Bulletin and Review, 9, 212-238. 505 Marti, S., Sigman, M., & Dehaene, S. (2012). A shared cortical bottleneck underlying 506 Attentional Blink and Psychological Refractory Period. Neuroimage, 59, 2883-507 2898. 508 Maslovat, D., Chua, R., Spencer, H. C., Forgaard, C. J., Carlsen, A. N., & Franks, I. M. 509 (2013). Evidence for a response preparation bottleneck during dual-task 510 Startle Refractory Period    25 performance: effect of a startling acoustic stimulus on the psychological 511 refractory period. Acta Psychologica, 144, 481-487. 512 Maslovat, D., Hodges, N. J., Chua, R., & Franks, I. M. (2011). Motor preparation and the 513 effects of practice: Evidence from startle. Behavioral Neuroscience, 125, 226-514 240. 515 Miller, J., & Reynolds, A. (2003). The locus of redundant-targets and nontargets effects: 516 evidence from the psychological refractory period paradigm. Journal of 517 Experimental Psychology: Human Perception and Performance, 29, 1126-1142. 518 Miller, J., & Ulrich, R. (2008). Bimanual response grouping in dual-task paradigms. 519 Quarterly Journal of Experimental Psychology, 61, 999-1019. 520 Pashler, H. E. (1994). Dual-task interference in simple tasks: data and theory. 521 Psychological Bulletin, 116, 220-244. 522 Pashler, H. E. (1998). The psychology of attention. Cambridge, MA: MIT Press. 523 Raymond, J. E., Shapiro, K. L., & Arnell, K. M. (1992). Temporary suppression of visual 524 processing in an RSVP task: an attentional blink? Journal of Experimental 525 Psychology: Human Perception and Performance, 18, 849-860. 526 Ruthruff, E., Johnston, J. C., Van Selst, M., Whitsell, S., & Remington, R. (2003). 527 Vanishing dual-task interference after practice: has the bottleneck been eliminated 528 or is it merely latent? Journal of Experimental Psychology: Human Perception 529 and Performance, 29, 280-289. 530 Startle Refractory Period    26 Smith, M. C. (1969). The effect of varying information on the psychological refractory 531 period. Acta Psychologica, 30, 220-231. 532 Stevenson, A. J., Chiu, C., Maslovat, D., Chua, R., Gick, B., Blouin, J. S., & Franks, I. 533 M. (2014). Cortical involvement in the StartReact effect. Neuroscience, 269, 21-534 34. 535 Telford, C. W. (1931). The refractory phase of voluntary and associative responses. 536 Journal of Experimental Psychology, 14, 1-36. 537 Thackray, R. I., & Touchstone, R. M. (1970). Recovery of motor performance following 538 startle. Perceptual and Motor Skills, 30, 279-292. 539 Thackray, R. I., Touchstone, R. M., & Jones, K. N. (1972). Effects of simulated sonic 540 booms on tracking performance and autonomic response. Aerospace Medicine, 541 43, 13-21. 542 Ulrich, R., Fernandez, S. R., Jentzsch, I., Rolke, B., Schroter, H., & Leuthold, H. (2006). 543 Motor limitation in dual-task processing under ballistic movement conditions. 544 Psychological Science, 17, 788-793. 545 Ulrich, R., & Miller, J. (2008). Response grouping in the psychological refractory period 546 (PRP) paradigm: models and contamination effects. Cognitive Psychology, 57, 547 75-121. 548 Valls-Solé, J., Kumru, H., & Kofler, M. (2008). Interaction between startle and voluntary 549 reactions in humans. Experimental Brain Research, 187, 497-507. 550 Startle Refractory Period    27 Valls-Solé, J., Rothwell, J. C., Goulart, F., Cossu, G., & Munoz, E. (1999). Patterned 551 ballistic movements triggered by a startle in healthy humans. Journal of 552 Physiology, 516.3, 931-938. 553 Vlasak, M. (1969). Effect of startle stimuli on performance. Aerospace Medicine, 40, 554 124-128. 555  556 557 Startle Refractory Period    28 Figure Captions 558 Figure 1. Mean verbal reaction time (RT1, with error bars representing 95% confidence 559 intervals) for various SOA intervals (top panel, A) and blocks (bottom panel, B), 560 separated by stimulus type (startle and non-startle trials). In the top panel, a single 561 asterisk (*) represent a main effect of stimulus, while a double asterisk (**) represent 562 longer RT1 in the single-task condition. In the bottom panel, the double asterisk (**) 563 represents a main effect of block, with decreasing RT1 with practice. 564 Figure 2. Mean key-lift reaction time (RT2, with error bars representing 95% confidence 565 intervals) for various SOA intervals, separated by stimulus type (startle and non-startle), 566 as compared to single-task performance (solid black line). Non-startle trials showed a 567 typical PRP effect in which shorter SOAs (100 ms, 200 ms and 500 ms) resulted in 568 significantly longer (**) RT2 latencies. In contrast to the predicted propagation effect, 569 significantly longer (*) RT2 latencies were found for startle trials at the 100 ms and 200 570 ms SOA conditions. 571 Figure 3. Schematic of predicted versus actual results. In the baseline (top) condition, 572 stimuli (S) are separated by a stimulus onset asynchrony (SOA). The shaded portion 573 represents the bottleneck portion of the task, which cannot start for task 2 until completed 574 for task 1. This results in a psychological refractory period (PRP) in which the second 575 response (R) has a delayed reaction time (RT). The current experiment replaced S1 with a 576 startling acoustic stimulus (SAS), resulting in a reduced RT1. The prediction of 577 propagation effects (middle panel) is that the reduction in RT1 is inherited by RT2. 578 However, actual results (bottom panel) showed an increase in RT2, which we attribute to 579 a startle refractory period (SRP). 580 Startle Refractory Period    29 Figure 4. Mean Reaction time (RT) differences between startle and non-startle trials for 581 various SOA intervals (significant differences are illustrated with an asterisk). Black bars 582 represent RT1 “savings” due to shorter latency verbal RT on startle trials while grey bars 583 represent RT2 delay due to longer latency key-lift RT on startle trials. These effects are 584 shown as cumulative as RT1 savings on startle trials were predicted to be propagated to 585 RT2 but instead RT2 values were longer for startle trials. 586 587 Startle Refractory Period    30 Figure 1 588  589  590 591 Startle Refractory Period    31 Figure 2 592  593  594  595 596 Startle Refractory Period    32 Figure 3 597  598  599  600 601 Startle Refractory Period    33 Figure 4 602  603  604 


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