Fig 1.
Illusory gap/overlap in the construction of saccade latencies.
Illustration of how multisensory integration (MSI) could introduce an illusory overlap/gap for negative/positive SOAs, by binding temporally the target onset with the beep signal (left/middle panel). This shift in time of the perceived target onset would produce an MSI theoretical latency that also shifts. However, the illusory overlap/gap would also produce an increase/decrease of saccade latencies that would go against the MSI theoretical effect and reduce the observed latency modulation. The right panel illustrates how adding a real overlap, arbitrarily set to half of the SOA, could compensate for the illusory gap introduced by the MSI for positive SOAs and thereby allowing a larger expression of the MSI modulation in the observed saccade latency.
Fig 2.
Time course of a trial illustrating the overlap (left) and gap (right) situations, the beep signal and the response mode for the experiments. A central fixation spot was presented for a random duration ranging from 750 to 1250 ms. The target appeared left/right at an eccentricity of 8°, either simultaneously (no-gap or step conditions), before (overlap conditions) or after the fixation offset (gap conditions). A short beep was played either before or after target onset (SOA ranging from −120 to +120 ms). In baseline conditions, no beep was played. Participants were instructed to shift their gaze towards the target (Experiments 1a, 2 and 3) or to indicate the target side by pressing the left/right mouse button (Experiment 1b), as fast and as accurately as possible. The response latency corresponded to the delay between target onset and the start of the saccade or the mouse-button press.
Fig 3.
Effect of gap compensation on saccade latency. (Left panel) Median latencies averaged across participants for the No gap and Gap compensated conditions as a function of SOA (top) and their corresponding nscores (bottom). Dashed lines show the No beep baseline condition level. Statistics performed on the nscores included a single sample t-test to compare the 0-ms SOA reference condition with the No beep condition (black arrow), and for each gap factor separately, paired t-tests compared this reference with each SOA condition (colored stars for each SOA above the X-axis). Finally, paired t-tests compared No gap and Gap compensated conditions for each SOA (black stars between curves). Three stars indicate highly significant differences for the paired t-tests after Bonferroni correction (p<0.00556). (Right panel) Modulation amplitude of latency median nscores (top) and slopes of the linear regressions within the range [−60 ms, +60 ms] (bottom) computed for each gap factor. Grey lines show individual results and error bars indicate inter-individual SEM. Comparisons were done using paired t-tests.
Fig 4.
Saccade latency (left) and manual reaction time (right) distributions. Distributions for the No beep condition and for the No gap and Gap compensated conditions plotted for each SOA, with data pooled from all participants.
Fig 5.
Effect of gap compensation on MRTs. (Left panel) Median latencies averaged across participants for the No gap and Gap compensated conditions as a function of SOA (top) and their corresponding nscores (bottom). Dashed lines show the No beep baseline condition level. Statistics performed on the nscores included a single sample t-test to compare the 0-ms SOA reference condition with the No beep condition (black arrow), and for each gap factor separately, paired t-tests compared this reference with each SOA condition (colored stars for each SOA above the X-axis). Finally, paired t-tests compared No gap and Gap compensated conditions for each SOA (black stars between curves). Three stars indicate highly significant differences for the paired t-tests after Bonferroni correction (p<0.00556). (Right panel) Modulation amplitude of latency median nscores (top) and slopes of the linear regressions within the range [−60 ms, +60 ms] (bottom) computed for each gap factor. Grey lines show individual results and error bars indicate inter-individual SEM.
Fig 6.
Effect of various gap compensation factors on saccade latency. (Left panel) Median latencies averaged across participants for each gap factor in {+½, 0, -½, -1} as a function of SOA (top) and their corresponding nscores (bottom). Dashed lines correspond to the No beep baseline condition. Statistics performed on the nscores included a single sample t-test to compare the 0-ms SOA reference condition with the No beep condition (black arrow), and for each gap factor separately, paired t-tests compared this reference with each SOA condition (colored stars for each SOA above the X-axis). Three stars indicate highly significant differences after Bonferroni correction (p<0.00294) while single stars and two stars indicate significant differences with less (p<0.005) or without correction (p<0.05). (Right panel) Modulation amplitude of the latency median nscores (top) and slopes of the linear regressions within the range [−60 ms, +60 ms] (bottom) computed for each gap factor. Grey lines show individual results and error bars indicate inter-individual SEM.
Fig 7.
Saccade latency distributions. Distributions for the No beep condition and for the different Gap factor conditions plotted for each SOA, with data pooled from all participants.
Fig 8.
Comparing the effects of beep MSI modulation and gap/overlap on saccade latency. (Left panel) Median latencies averaged across participants for the Beep conditions (no gap/overlap) and the Gap conditions (no beep) as a function of SOA or gap duration (top), and their corresponding nscores (bottom). Statistics performed for each group of conditions separately on the nscores, included paired t-tests comparing each SOA or gap condition with the 0-ms SOA or Gap duration reference conditions, respectively (colored stars above the X-axis). Three stars indicate highly significant differences after Bonferroni correction (p<0.0125) while single stars indicate significant differences without correction (p<0.05). (Right panel) Saccade latency distributions for the Beep conditions and the Gap conditions plotted for each SOA and Gap, respectively, with data pooled from all participants.
Fig 9.
Testing linear combination of beep MSI modulation and gap/overlap on saccade latency. Comparison of the Gap compensated condition from Experiment 1a for the same participants, with the combination of the Beep conditions and Gap conditions from Experiment 3, using the same baseline condition (respectively No beep and No gap).