東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Integration of local features for th...

Clark, Andrew Michael.

Linked to FindBook

Google Book

Amazon

博客來

Integration of local features for the computation of pattern velocity in Macaque middle temporal cortical neurons.

Record Type:	Language materials, printed : Monograph/item
Title/Author:	Integration of local features for the computation of pattern velocity in Macaque middle temporal cortical neurons./
Author:	Clark, Andrew Michael.
Description:	138 p.
Notes:	Adviser: David C. Bradley.
Contained By:	Dissertation Abstracts International68-10B.
Subject:	Biology, Neuroscience. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3287033
ISBN:	9780549296447

Integration of local features for the computation of pattern velocity in Macaque middle temporal cortical neurons.
Clark, Andrew Michael.

Integration of local features for the computation of pattern velocity in Macaque middle temporal cortical neurons. - 138 p.

Adviser: David C. Bradley.

Thesis (Ph.D.)--The University of Chicago, 2007.

Estimating the motion of objects and surfaces given only the time-varying distribution of light intensities impinging upon the retinea is a difficult problem. Computational solutions to this problem often assume a hierarchical form. At the first stage motion is detected point-by-point within the scene, providing a high resolution representation of velocity. As these estimates are ambiguous, that is, a family of object motions can result in identical local samples, a second, integration, stage is required to smooth out noise and recover object velocities. While psychophysical studies have marshaled a preponderance of evidence for such a two-stage scheme in human vision, physiological evidence has proven harder to obtain. Early reports identified the primary visual (V1) and middle temporal (MT) cortical areas in the macaque monkey as the putative neural loci of these detection and integration stages, respectively. However, more recent studies have called this segregation into question. Separate reports have alternatively suggested that signals from V1 neurons do provide unambiguous estimates of motion, and that MT neurons only integrate local motion signals when they are spatially overlapping; a special and highly restrictive case. In an effort to reconcile these disparate physiological results and relate them to psychophysical accounts of motion integration we recorded the responses of single MT neurons to several classes of stimuli that were specifically designed to probe the following set of questions. (i) Do MT neurons integrate local motion signals when they are spatially localized? (ii) Can any differences between MT responses in spatially overlapping and non-overlapping conditions be attributed to the inclusion of second-order motions in the overlapping condition? (iii) Which class of computational models best predicts the manner in which MT neurons seem to integrate local motion signals? Our results indicate that MT neurons do integrate local motion signals when they are spatially distributed, and suggest that any differences between spatial conditions can be, at least partially, attributable to sensitivity to the second-order signals in the spatially overlapping case. Finally, the models that best explains these findings are those in which MT neurons perform a weighted average of the local stimulus velocities in their receptive field.

ISBN: 9780549296447Subjects--Topical Terms:

1017680
Biology, Neuroscience.

Integration of local features for the computation of pattern velocity in Macaque middle temporal cortical neurons.
LDR:03328nam 2200289 a 45 001 943704
005 20110520
008 110520s2007 ||||||||||||||||| ||eng d
020 $a 9780549296447
035 $a (UMI)AAI3287033
035 $a AAI3287033
040 $a UMI $c UMI
100 1 $a Clark, Andrew Michael. $3 1267741
245 1 0 $a Integration of local features for the computation of pattern velocity in Macaque middle temporal cortical neurons.
300 $a 138 p.
500 $a Adviser: David C. Bradley.
500 $a Source: Dissertation Abstracts International, Volume: 68-10, Section: B, page: 6501.
502 $a Thesis (Ph.D.)--The University of Chicago, 2007.
520 $a Estimating the motion of objects and surfaces given only the time-varying distribution of light intensities impinging upon the retinea is a difficult problem. Computational solutions to this problem often assume a hierarchical form. At the first stage motion is detected point-by-point within the scene, providing a high resolution representation of velocity. As these estimates are ambiguous, that is, a family of object motions can result in identical local samples, a second, integration, stage is required to smooth out noise and recover object velocities. While psychophysical studies have marshaled a preponderance of evidence for such a two-stage scheme in human vision, physiological evidence has proven harder to obtain. Early reports identified the primary visual (V1) and middle temporal (MT) cortical areas in the macaque monkey as the putative neural loci of these detection and integration stages, respectively. However, more recent studies have called this segregation into question. Separate reports have alternatively suggested that signals from V1 neurons do provide unambiguous estimates of motion, and that MT neurons only integrate local motion signals when they are spatially overlapping; a special and highly restrictive case. In an effort to reconcile these disparate physiological results and relate them to psychophysical accounts of motion integration we recorded the responses of single MT neurons to several classes of stimuli that were specifically designed to probe the following set of questions. (i) Do MT neurons integrate local motion signals when they are spatially localized? (ii) Can any differences between MT responses in spatially overlapping and non-overlapping conditions be attributed to the inclusion of second-order motions in the overlapping condition? (iii) Which class of computational models best predicts the manner in which MT neurons seem to integrate local motion signals? Our results indicate that MT neurons do integrate local motion signals when they are spatially distributed, and suggest that any differences between spatial conditions can be, at least partially, attributable to sensitivity to the second-order signals in the spatially overlapping case. Finally, the models that best explains these findings are those in which MT neurons perform a weighted average of the local stimulus velocities in their receptive field.
590 $a School code: 0330.
650 4 $a Biology, Neuroscience. $3 1017680
650 4 $a Biology, Physiology. $3 1017816
650 4 $a Psychology, Physiological. $3 1017869
690 $a 0317
690 $a 0719
690 $a 0989
710 2 $a The University of Chicago. $3 1017389
773 0 $t Dissertation Abstracts International $g 68-10B.
790 $a 0330
790 1 0 $a Bradley, David C., $e advisor
791 $a Ph.D.
792 $a 2007
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3287033