2019\u5e7410\u670816\u65e5\uff08\u6c34\uff09
16:00-17:00
\u4f1a\u5834\uff1a CiNet\u30001F\u3000\u5927\u4f1a\u8b70\u5ba4<\/p>\n\n\n\n
\u201cDynamics and mechanisms of flexible information processing in the visual system\u201d<\/p>\n\n\n\n
Udo Ernst
University of Bremen<\/p>\n\n\n\n
Abstract:
Computation in the visual system must be flexible, for rapidly configuring processing towards changing behavioral goals, and for reacting uickly to dynamic changes in our environment. Our lab is interested in quantifying the dynamics and efficacy of flexible information processing, and in investigating putative neural mechanisms underlying flexible computation. For this purpose we combine theoretical work and modelling with analysis of electrophysiological recordings in awake behaving macaque monkeys.<\/p>\n\n\n\n
In this presentation we would like to focus on two specific aspects of flexible visual processing: Attentional modulation of fast transient neural responses in area MT to stimulus changes, and selective processing of visual stimuli by the collective dynamics of neural populations in areas V1\/V2 and V4.<\/p>\n\n\n\n
In particular, we will show that selective processing is realized by transferring an attended visual signal in pulsed packages during excitability peaks of gamma activity in area V4. To explain this experimental result we constructed a model which predicts that such a selective transfer can be realized by raising the firing rate of V1\/V2 populations carrying the attended (target) signal above the firing rates of V1\/V2 populations representing the non-attended (distractor) signal. Surprisingly, recent experiments confirm this prediction even for situations in which the target has a much lower luminance contrast than the distractor. This leads to unexpectedly large attentional rate modulations for V1\/V2 neurons (40%-60%) which were not observed in previous studies.<\/p>\n\n\n\n
Furthermore, we investigated transient responses in area MT which exhibit a rich and previously unexplained dynamics depending non-linearly on the magnitude of a stimulus change. We show that a simple dynamic divisive normalization model is able to quantitatively explain peak height and shape of the transient response. Specifically, it predicts that attention will always increase the slope of the transient, irrespectively of the current activation state of the neuron. This prediction was confirmed for both positive and negative transients by recent experiments performed by our collaborators. Interestingly, the same model is also capable to explain the large rate modulations in V1\/V2, thus hinting towards a universal control mechanism by which attention flexibly configures neural processing in parallel between and within visual areas.<\/p>\n\n\n\n
\u62c5\u5f53\uff1a\u00a0\u7530\u6751 \u5f18\uff08\u85e4\u7530\u30b0\u30eb\u30fc\u30d7\uff09<\/p>\n","protected":false},"featured_media":0,"template":"","acf":[],"_links":{"self":[{"href":"http:\/\/cinetjp-static3.nict.go.jp\/japanese\/wp-json\/wp\/v2\/event\/1509"}],"collection":[{"href":"http:\/\/cinetjp-static3.nict.go.jp\/japanese\/wp-json\/wp\/v2\/event"}],"about":[{"href":"http:\/\/cinetjp-static3.nict.go.jp\/japanese\/wp-json\/wp\/v2\/types\/event"}],"wp:attachment":[{"href":"http:\/\/cinetjp-static3.nict.go.jp\/japanese\/wp-json\/wp\/v2\/media?parent=1509"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}