The neural substrates of object-ensemble processing in the human brain

人脑中对象集成处理的神经基础

• 批准号: 435647-2013
• 负责人: Cant, Jonathan
• 金额: $ 2.11万
• 依托单位: University of Toronto
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2015
• 资助国家: 加拿大
• 起止时间: 2015-01-01 至 2016-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=574136
• 关键词: neural; substrates; object; ensemble; processing

The way in which we see the world is extremely important to our survival as human beings. For example, the ability to be able to find food to eat, friends to socialize with, and partners to mate with requires a visual system that can rapidly and accurately identify the objects and individuals in any given environment. Research in Dr. Jonathan Cant's laboratory explores these issues, in particular how the human brain processes objects and scenes. Previous research using a modern brain imaging technique known as functional magnetic resonance imaging (fMRI) has demonstrated that a region of the brain known as the parahippocampal place area (or PPA for short) plays a large role in scene perception, and does so by processing the 3D spatial structure, or geometry, of scenes. Dr. Cant has recently discovered however, that the PPA does more than just process the geometry of a scene. Specifically, he was the first to show that the PPA is involved in processing what is known as large collections, or ensembles, of multiple objects (e.g. all the leaves on a tree). While we do not know much about how the brain processes object ensembles, we do know that they are extremely important, as ensemble processing entails collapsing a lot of redundant visual information into a single summary feature (e.g. the average size of all the leaves on a tree), thereby saving the brain a lot of time and energy. We would simply not be able to perceive scenes efficiently if our visual system had to process every individual leaf on a tree one-by-one, or every individual blade of grass in a field! Thus, the goal of Dr. Cant's current research program is to understand how the brain processes ensembles of multiple objects, and how this type of processing relates to our ability to pick out and identify one object within the ensemble (that is, single-object perception). This research will not only help scientists understand how the visual brain is organized, but will help engineers develop better machine-vision systems, since many of the design principles used in computer vision systems are inspired by research on how humans perceive individual objects within complex scenes (e.g. the ability of digital cameras to recognize where people are standing in the scene you are taking a picture of).

我们看待世界的方式对我们人类的生存至关重要。例如，寻找食物、社交朋友和交配伙伴的能力，需要一种能够在任何给定环境中快速准确地识别物体和个体的视觉系统。乔纳森·坎特博士实验室的研究探索了这些问题，特别是人类大脑如何处理物体和场景。先前的研究使用了一种被称为功能磁共振成像（fMRI）的现代脑成像技术，已经证明了大脑中一个被称为海马体旁区（简称PPA）的区域在场景感知中起着重要作用，它通过处理场景的3D空间结构或几何形状来实现。然而，坎特博士最近发现，PPA不仅仅是处理场景的几何形状。具体来说，他是第一个表明PPA参与处理所谓的大型集合或多个对象的集合（例如树上的所有叶子）的人。虽然我们不太了解大脑如何处理对象集合，但我们知道它们非常重要，因为集合处理需要将大量冗余的视觉信息折叠成单个汇总特征（例如，树上所有叶子的平均大小），从而为大脑节省了大量时间和精力。如果我们的视觉系统必须一个接一个地处理树上的每一片叶子，或者田野里的每一片草，我们就无法有效地感知场景！因此，坎特博士目前研究项目的目标是了解大脑如何处理多个物体的集合，以及这种处理方式与我们在集合中挑选和识别一个物体的能力（即单物体感知）之间的关系。这项研究不仅将帮助科学家了解视觉大脑是如何组织的，而且将帮助工程师开发更好的机器视觉系统，因为计算机视觉系统中使用的许多设计原则都受到了人类如何在复杂场景中感知单个物体的研究的启发（例如，数码相机能够识别你正在拍照的场景中站在哪里的人）。