【Talk1】
講者：Michèle Fabre-Thorpe, Ph.D. （Centre de Recherche Cerveau et Cognition (CerCo) CNRS et Université Toulouse3 - Toulouse）
演講題目：Natural scenes : Early interactions in object and context processing flows.
摘要：
The efficiency of the primate visual system at recognizing hundreds of thousands of objects in complex and rich environments is astonishing. When required to categorize animals, faces, vehicles… in natural photographs flashed for only 20ms, humans and monkeys are fast and very accurate even in extreme conditions of presentation (lack of color, low contrasts, strong masking…).
One of the main characteristics of early visual processing is its massive parallelism. When subjects are required to perform an animal/non-animal categorization task, two scenes can be processed as fast as one! When using natural photographs as stimuli, objects are embedded in their natural settings and such parallelism should lead to object/context interactions. The temporal dynamics of such interactions is still source of controversy. Object and scene gist could interact early, late, or even not at all! The extraction of scene gist statistics might even be sufficient in many cases to conclude on the presence/absence of a given object. The extent of object/context interactions will be shown in a series of experiment using fast manual categorization of natural scenes. The minimal latency at which such interactions could be evidenced was assess at 160 ms by the use of a forced choice saccadic categorization task that made possible the exploration of a very early temporal window of visual processing.
A model will be presented in which object and contextual information interact in a feed-forward ascending flow of processing. Interference and facilitation are determined by the frequency with which populations of object selective neurons are co-activated in the processing of our surrounding world.
【Talk2】
講者：Simon Thorpe, Ph.D. （CNRS Brain and Cognition Research Centre, Université de Toulouse III, Toulouse, France）
演講題目：How can our brains store visual (and auditory) memories that last a lifetime?
摘要：
People in their 50s and 60s can recognize images and sounds that they have not re-experienced for several decades. How does the brain manage to keep these memory traces intact? I will describe some experimental and modeling work that supports the radical suggestion that these extremely long term memories involve the formation of highly selective neurons, tuned to stimulus patterns that were presented repeatedly at some time in the past - effectively “grandmother cells”. Such highly selective neurons can be produced using a simple learning rule based on Spike Time Dependent Plasticity (STDP) that leads neurons to become selective to spatiotemporal patterns of input spikes that occur repeatedly. Even more radical is the suggestion that the neocortex may contain a substantial proportion of totally silent neurons - a sort of neocortical dark matter. Such neurons will effectively remain silent until the original trigger stimulus is shown again. If STDP-like rules apply, the absence of firing would mean that the sets of synaptic weights could remain intact virtually indefinitely.