2016 Sep 29

CBB Seminar - Rick Born (Harvard)

Date: 

Thursday, September 29, 2016, 12:00pm

Location: 

WJH 765

Bottom-up and top-down inputs drive the variability of cortical neurons

Neurons in the cerebral cortex respond inconsistently to a repeated sensory stimulus, so how can they provide the basis for stable sensory experiences? Although the exact causes of neuronal response variability are unknown, the consistency with which it has been observed across a variety of cortical regions has encouraged the general view that each cell produces random spike patterns that noisily represent its response rate. In contrast to this view, we discovered that reversibly inactivating sources of either bottom-up (V2-to- MT) or top-down (V2-to-V1) input to cortical visual areas in the alert primate reduced both the spike train irregularity and the trial-to- trial variability of single neurons. A simple network model of integrate-and- fire neurons in which a fraction of the pre-synaptic inputs are silenced can reproduce this reduction in variability, provided that there exist temporal correlations primarily within, but not between, excitatory and inhibitory input pools. A large component of the variability of cortical neurons can therefore be ascribed to synchronous input produced by signals arriving from multiple sources. Taken together, our results impose strong constraints on theories of neuronal variability by causally linking the presence of bottom-up and top-down input to the spiking statistics of cortical neurons.

2016 Oct 06

CBB Seminar - Jan Drugowitsch (Harvard)

Date: 

Thursday, October 6, 2016, 12:00pm

Location: 

WJH 765

Behavioral variability as a signature of approximate mental computations

The variability of human decisions to perceptual evidence is
frequently larger than what can be explained by variability in the
evidence itself. This additional variability has been hypothesized to
arise from noisy sensory processing or stochastic action selection,
thus placing its origin at the peripheries of the decision-making
process. Based on a combination of psychophysical experiments and
computational modeling I will provide evidence that most of the
variability instead has its origin in imprecisions in central
probabilistic mental computations. This finding supports the view that
real-world decisions are computationally intractable and need to be
approximated, causing this additional variability. Indeed, a sizable
fraction of the variability we observed in our experiments can be
attributed to deterministic deviations from Bayes-optimal decision
strategies. The residual unstructured variability seems to arise from
perturbed temporal evidence accumulation, but its precise origin still
remains to be determined.