|
Pinpointing the origin of behavioural variability in perceptual decision-making| old_uid | 12303 |
|---|
| title | Pinpointing the origin of behavioural variability in perceptual decision-making |
|---|
| start_date | 2013/04/08 |
|---|
| schedule | 11h |
|---|
| online | no |
|---|
| summary | Human perceptual decisions often vary beyond what can be explained by variability in the underlying sensory evidence. It has been hypothesised that this "intrinsic" variability stems either from noise in the inference process (due to suboptimal computations) or from noise in the selection process (by drawing random samples from a posterior belief), but the relative contributions of these two noise sources remain unclear. Using a combination of psychophysics and behavioural modelling, I will first show that inference noise predominates over selection noise under stable sensorimotor contingencies. I will then characterise two constraints on sequential information processing which typically produce inference noise in perceptual decisions. For this purpose, I will use model-based analyses of human EEG signals recorded during multi-sample perceptual categorisation. The findings emphasise the existence of a central processing bottleneck which limits, in terms of capacity, our ability to integrate perceptual information over space and time. |
|---|
| responsibles | Blancho |
|---|
| |
|