Subjective time is predicted by local and early visual processing
Curation statements for this article:-
Curated by eLife
Summary: This work uses electro-encephalographic (EEG) recordings combined with an interesting experimental approach to measure temporal expansion/compression. Specifically, the question addressed here is whether adaptation to visual motion affects perceived duration, and if so, how spatially confined these effects are with respect to the processing of the stimulus in early visual areas. The authors find consistent evidence that a visual reference is judged as shorter/longer depending on a previous adaptation. They report several EEG analyses suggesting the early visual activity is correlated with such temporal distortions. This manuscript is of potential interest to cognitive neuroscientists specifically interested in temporal aspects of visual processing and time perception. Although the paradigm is well suited to assess the authors' question, the behavioral data as well as the electrophysiological analyses show important shortcomings currently hindering the interpretation of the results, and necessitating substantial revisions to the current work. Additionally, further methodological details are required to strengthen the manuscript.
This article has been Reviewed by the following groups
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
- Evaluated articles (eLife)
Abstract
Time is as pervasive as it is elusive to study, and how the brain keeps track of millisecond time is still unclear. Here we studied the mechanisms underlying duration perception by looking for a neural signature of subjective time distortion induced by motion adaptation. We recorded electroencephalographic signals in human participants while they were asked to discriminate the duration of visual stimuli after translational motion adaptation. Our results show that distortions of subjective time can be predicted by the amplitude of the N200 event-related potential and by the activity in the Beta band frequency spectrum. Both effects were observed from occipital electrodes contralateral to the adapted stimulus. Finally, a multivariate decoding analysis highlights the impact of motion adaptation throughout the visual stream. Overall, our findings show the crucial involvement of local and low-level perceptual processes in generating a subjective sense of time.
Article activity feed
-
Reviewer #2:
This work combines an interesting experimental approach to measure temporal expansion/compression with EEG recordings. The authors find consistent evidence that a visual reference is judged as shorter/longer dependent on a previous adaptation. They report several EEG analyses suggesting the early visual activity is correlated with such temporal distortions.
Strengths:
The paper uses an interesting design to try to isolate temporal compression/expansion. The behavioral results are consistent and they show several different EEG analyses. The main result, of beta power being correlated with temporal processing, is consistent with previous reports.
Weaknesses:
The paper would strongly benefit from more details on some of the methodologies and results. In several moments, the authors show measures that are subtracted or …
Reviewer #2:
This work combines an interesting experimental approach to measure temporal expansion/compression with EEG recordings. The authors find consistent evidence that a visual reference is judged as shorter/longer dependent on a previous adaptation. They report several EEG analyses suggesting the early visual activity is correlated with such temporal distortions.
Strengths:
The paper uses an interesting design to try to isolate temporal compression/expansion. The behavioral results are consistent and they show several different EEG analyses. The main result, of beta power being correlated with temporal processing, is consistent with previous reports.
Weaknesses:
The paper would strongly benefit from more details on some of the methodologies and results. In several moments, the authors show measures that are subtracted or normalized based on other conditions. Although these normalizations can sometimes help to illustrate effects, it also makes it harder to understand the data in a more general sense. For example, in their behavioral results, the authors present an Adaptation Effect to quantify temporal compression/expansion. It would also help if authors present the raw estimates of Points of Subjective Equality across all conditions (including the unadapted condition) so that the reader can have a better understanding of the effects. It would be even better if the average proportion of responses for each duration was shown so that readers can see differences in PSE, JND, and guess/lapse rates.
Further details about the EEG analysis would also help the readers. For example, it is not totally clear how the FFT analysis was performed. It would be important to add information about whether data was analyzed using moving windows, the size of the windows, whether there was an overlap between windows, whether there was a baseline correction and what was the baseline.
Several of the conclusions of the authors are based on linear mixed effect (LME) regressions in which the PSE or the behavioral effect is the dependent variable and an EEG measure is used as one of the fixed effects. However, in some of the analysis, it is not really clear how this was performed (for example, whether this was done at the single-trial or at the averaged data). Critically, it would help the reader if more output (both tables and graphs) were shown for these analyses so that what is being analyzed and concluded is made clearer.
-
Reviewer #1:
The question is interesting, and the paradigm in principle well suited to answer it. Unfortunately, a number of shortcomings hinder a clear interpretation of the results. I think that the paper, notably the EEG analyses, need to be revised substantially, which might affect the results. Therefore I will just list the main points which need to be addressed and not go in more detail.
The behavioral effect of adaptation on duration perception appears very unspecific, namely it occurs in all but the spatially neutral condition. The authors conclude that the inversely directed motion did not have an effect because it did not survive the Bonferroni correction, yet they report a p-value of 0.02 and Cohen's d of 0.58, suggesting a medium effect. In order to prove the absence of an effect, I suggest to report Bayes factors, and only …
Reviewer #1:
The question is interesting, and the paradigm in principle well suited to answer it. Unfortunately, a number of shortcomings hinder a clear interpretation of the results. I think that the paper, notably the EEG analyses, need to be revised substantially, which might affect the results. Therefore I will just list the main points which need to be addressed and not go in more detail.
The behavioral effect of adaptation on duration perception appears very unspecific, namely it occurs in all but the spatially neutral condition. The authors conclude that the inversely directed motion did not have an effect because it did not survive the Bonferroni correction, yet they report a p-value of 0.02 and Cohen's d of 0.58, suggesting a medium effect. In order to prove the absence of an effect, I suggest to report Bayes factors, and only interpret the effect as absent if the Bayes factor is conclusive towards the H0.
In my view, if there was an effect of inversely directed motion, this poses a question as to the successful demonstration of specific adaptation effects in the behavior, which needs to be taken into account in the interpretation.
The EEG analyses and displayed results show some important shortcomings, which hinder a clear interpretation at this stage. Just to list a few main points:
-As apparent from Figures 3-5, the time-frequency plots show a lot of stripes and pixels, when one would expect rather smooth transitions over frequency and time. This suggests that the parameters for the time-frequency transformation might not be appropriate.
-The analyses compare time windows that differ in many respects, for instance the 15 s long adaptation phase versus short-lived stimulus-evoked activity at reference onset. Interpreting these differences as specific to the duration distortion effects does not seem justified, due to the diverging inputs presented during those time windows.
-Important aspects of the paradigm are not taken into account in the EEG analyses, for instance the fact that participants perform a saccade between the offset of adaptation and the onset of the reference. The saccade-related signatures in the EEG have to be accounted or controlled for, especially for effects occurring after adaptation offset.
-Some of the effects (for instance the decoding analysis, or the linear mixed models testing for additive but not interactive effects) show differences in EEG activity related to visual processing of the stimuli, but might not specifically relate to the duration distortions. In my view, more trivial differences in processing the visual inputs should be accounted for (see also the point above), and clearly separated from specific timing effects.
-
Summary: This work uses electro-encephalographic (EEG) recordings combined with an interesting experimental approach to measure temporal expansion/compression. Specifically, the question addressed here is whether adaptation to visual motion affects perceived duration, and if so, how spatially confined these effects are with respect to the processing of the stimulus in early visual areas. The authors find consistent evidence that a visual reference is judged as shorter/longer depending on a previous adaptation. They report several EEG analyses suggesting the early visual activity is correlated with such temporal distortions. This manuscript is of potential interest to cognitive neuroscientists specifically interested in temporal aspects of visual processing and time perception. Although the paradigm is well suited to assess the authors' …
Summary: This work uses electro-encephalographic (EEG) recordings combined with an interesting experimental approach to measure temporal expansion/compression. Specifically, the question addressed here is whether adaptation to visual motion affects perceived duration, and if so, how spatially confined these effects are with respect to the processing of the stimulus in early visual areas. The authors find consistent evidence that a visual reference is judged as shorter/longer depending on a previous adaptation. They report several EEG analyses suggesting the early visual activity is correlated with such temporal distortions. This manuscript is of potential interest to cognitive neuroscientists specifically interested in temporal aspects of visual processing and time perception. Although the paradigm is well suited to assess the authors' question, the behavioral data as well as the electrophysiological analyses show important shortcomings currently hindering the interpretation of the results, and necessitating substantial revisions to the current work. Additionally, further methodological details are required to strengthen the manuscript.
-
