Symposium
Towards understanding how objects are perceived in our brain
シンポジウム
物体認識の脳内機構へのアプローチの最前線 |
|---|
| 7月26日（金）15:15～15:40 第2会場（朱鷺メッセ 2F メインホールA）
2S02a-1
Joint encoding of shape and surface properties in mid-level ventral visual cortex
Anitha Pasupathy（Pasupathy Anitha）1,2，Taekjun Kim（Kim Taekjun）1,2，Dina V Popovkina（Popovkina Dina V）3，Wyeth Bair（Bair Wyeth）1,2
1Dept. of Biological Structure, University of Washington
2Washington National Primate Research Center
3Dept. of Psychology, University of Washington
Recognizing a myriad visual objects rapidly is a hallmark of the primate visual system. Traditional theories of object recognition have focused on how critical form features, e.g. the orientation of edges or surfaces and the position of vertices, may be extracted from the visual image and utilized to recognize objects. While this idea is supported by neurophysiological studies demonstrating selectivity for the orientation of edges in early visual cortex and curvature of object boundaries in mid-level stages, it downplays the role of surface properties, e.g. surface color and texture, critical for scene understanding and object recognition in natural scenes. Furthermore, it ignores the fact that identifying the bounding contours of objects from an edge map can be extremely difficult. An alternative view argues that much of the processing in early and mid-level stages is focused on encoding surface characteristics of an image, as opposed to the orientation of edges. But it is unclear whether this strategy is compatible with the ability to recognize object form regardless of surface characteristics. Here we discuss neurophysiological evidence in support of the hypothesis that neurons in area V4, an intermediate stage in the ventral visual pathway, jointly encode both form and surface properties. Such a representation would be advantageous for segmenting objects from the background in natural scenes and simultaneously facilitate the recognition of form regardless of texture. Future experiments are necessary to determine how a joint representation in the mid-level emerges from representations in terms of local orientation and spatial frequency in V1. | | 7月26日（金）15:40～16:05 第2会場（朱鷺メッセ 2F メインホールA）
2S02a-2
皮質脳波法で読み解く物体視・記憶の大脳分散表現
Isao Hasegawa（長谷川 功）
新潟大院医歯総合研神経生理
The cerebral cortex is an ultimate complex system. Human neuroimaging studies and single-unit recording studies in non-human animals have emphasized functional localization of different aspects of object vision/memory in particular cortical subdivisions where activity of single neurons plays important roles. On the other hand, any cognitive function is associated with the spatial and temporal interactions of millions of neurons across a large-scale brain network. Even the simplest visual stimulation progressively activates intra- and inter-columnar neuronal interactions in the distributed network spanning the occipital, temporal, parietal, and frontal cortices, together with subcortical structures. Therefore, visualization of the dynamic patterns of synchronous neuronal activity along the surface of the cerebral gyri/sulci in various scales is critical to understanding how the brain perceives and recalls objects. We have developed, with micro-electro-mechanical systems, a mesh electrocorticographic (ECoG) electrode array for capturing propagation and distribution of synchronous neuronal activity in the sensory and association cortical areas in experimental animals. The electrode array is sufficiently flexible to allow minimally-invasive surgical implantation even into the macaque superior temporal sulcus. The inter-electrode space, freely scalable from 0.1 to 2.5 millimeter, allows penetration of metal microelectrodes or optrodes for simultaneous recording/stimulation. High-density ECoG recording simultaneously with multi-unit activity and local field potentials revealed how basic visual category and finer category information is distributed in spatiotemporal patterns of neuronal activity in the macaque anterior inferior temporal cortex. Area-wide ECoG recording spanning the macaque medial temporal lobe visualized distribution and reorganization of cortical activity maps through associative memory learning. These results demonstrate the feasibility of intensive and extensive recordings with flexible mesh electrodes in capturing signal flow/distribution across the cortical networks fundamental to object recognition. | | 7月26日（金）16:05～16:30 第2会場（朱鷺メッセ 2F メインホールA）
2S02a-3
How are priors and likelihood combined during object recognition?
Floris P de Lange（de Lange Floris P）
Radboud University Nijmegen, Netherlands
Perception and perceptual decision-making are strongly facilitated by prior knowledge about the probabilistic structure of the world. While the computational benefits of using prior expectation in perception are clear, there are myriad ways in which this computation can be realized. In my talk, I will discuss recent observations from my lab into how this process may be neurally implemented. Specifically, I will: 1) provide evidence for the separation of priors and likelihood in different layers of the cortex, by capitalizing on ultra-high field fMRI at 0.8 mm resolution in humans; 2) I will examine the integration of priors and likelihood in the context of object recognition in the visual ventral stream, ranging from the primary visual to medial enthorhinal cortex; 3) I will show that attention mediates this balance by increasing, rather than reducing, the effect of priors. These data may inform and constrain computational frameworks that implement probabilistic integration of priors and likelihood in perception, such as predictive coding.
References:
* de Lange FP, Heilbron M, Kok P (2018). How Do Expectations Shape Perception? Trends in Cognitive Sciences 22, 764-779.
* Summerfield C, de Lange FP (2014). Expectation in perceptual decision-making: neural and computational mechanisms. Nature Reviews Neuroscience 15, 745-756.
* Lawrence SJD, van Mourik T, Kok P, Koopmans PJ, Norris DG, de Lange FP (2018). Laminar Organization of Working Memory Signals in Human Visual Cortex. Current Biology 28,3435-3440. | | 7月26日（金）16:30～16:55 第2会場（朱鷺メッセ 2F メインホールA）
2S02a-4
Color: a tool to understand the organization and operation of the ventral visual pathway
Bevil Conway（Conway Bevil）
National Institutes of Health (USA)
An important job of the brain is to turn noisy sense data into representations that can drive action. My lab exploits color as a tool to understand this fundamental process. What is color for and how are color operations implemented in the brain? I will take up these questions, drawing upon neurophysiological recordings in macaque monkeys, fMRI in humans and monkeys, psychophysics, and color-naming in a non-industrialized Amazonian culture. My talk will have three parts. First, I will discuss results showing that the neural implementation of color depends on a multi-stage network that encodes color initially through a process that reflects the color statistics of the environment, and ultimately as a more uniform representation of color space within a mid-level stage in visual processing. Second, I will describe work suggesting that color is decoded by a series of stages within inferior temporal cortex and prefrontal cortex (PFC). In a surprising twist, these discoveries reveal a general principle for the organization and operation of inferior temporal cortex and provide evidence for a stimulus-driven functional organization of PFC. In the final part, I will describe two recent discoveries prompted by our neurobiological discoveries: a new interaction of color and face perception; and a universal pattern in color naming that reflects the color statistics of those parts of the world that we especially care about (objects). Together, the work supports the provocative idea that that basic color categories are an emergent property arising from the needs we place on the brain (including object recognition and the assignment of object valence), rather than a constraint determined by color encoding. | |
|
|
