« ATOM & When Pathway | トップページ | VSS 終了 »

VSS 途中経過

いま、フロリダのVSSという学会に来ている。
毎年来ているが、今年はどうもレベルが高いようだ。もしかしたら、自分が他人の研究をみる能力が向上したのかもしれない。あと、自分の実験が手法的にmultimodalになってきて、気になることが増えたというのもあるかもしれない。

学会の最初から来ていたわけではないからすべてを見てはいないが、面白かったもののアブストラクトをメモっておく。

11:45 207 Retinotopy versus category specificity throughout
primate cerebral cortex
Reza Rajimehr1 (reza@nmr.mgh.harvard.edu), Wim Vanduffel1, Roger Tootell1;
1NMR Athinoula A. Martinos Center, Massachusetts General Hospital (MGH),
Charlestown, MA
In initial stages of the visual hierarchy, stimuli are represented in point-topoint
(retinotopic) maps. However at higher stages in the ventral stream, it has been proposed that the retinotopic map is converted into a categoryselective cortical architecture (e.g. face representations in IT cortex of humans and macaques). To independently map and compare these two dimensions, we acquired fMRI from two awake monkeys and six human subjects, while presenting stimuli consisting of face- or place-based images, confined to retinotopically-specific apertures, during continuous central fixation. This common data was analyzed in two different ways: 1) based on category (faces vs. places), and 2) based on retinotopic location (e.g. foveal vs. peripheral stimuli). In macaques, the category comparison revealed face-selective regions in the posterior IT/STS, anterior IT, lateral parietal and inferior frontal cortex, consistent with previous studies. In human subjects, analogous face-selective regions were revealed in two major areas described previously: the fusiform gyrus (including FFA), and lateral occipital cortex (~V4d) - plus a smaller region in anterior temporal lobe, perhaps homologous to the anterior temporal face patch in macaques. The retinotopy analysis demonstrated a surprising degree of retinotopic organization in IT cortex. In both human and macaque subjects, each face-selective region typically contained subdivisions selective for either foveal or peripheral stimuli, or both. Retinotopic maps (polar and eccentricity) were also revealed in more than 20 additional areas in macaques, including frontal and parietal cortex, association cortex, and ventral occipital regions not described previously. The overall findings suggest a conjoined representation of object and space selectivity in the ventral pathway - not a ‘pure’ representation of object category.

そこまですごいわけではないかもしれないけど、ファンダメンタルで大事な実験をしているという印象だ。

4:45 224 Spatial Limits of Face Processing: Evidence from Face Aftereffects.
Seyed-Reza Afraz1 (afraz@fas.harvard.edu), Patrick Cavanagh1; 1Department of Psychology, Harvard University
We examined the spatial limits of face processing by measuring the spatial spread of the face aftereffect (FAE) following adaptation to a single face and following simultaneous adaptation to a face and its anti-face at different spatial separations. In the adaptation phase, one of the following sets of stimuli was presented with each stimulus placed at locations around a 3 deg circle centered at fixation: 1) a single face; 2) a face and its anti-face on opposite sides of fixation; 3) two faces and two anti-faces evenly spaced; and 4) one to four ellipses evenly spaced with the average size and color of face stimuli (non-adapted condition). In the test phase, a face stimulus with various morphing levels between the face and the anti-face was presented at a random location around the display circle and subjects had to report whether the test stimulus was seen as face or anti-face. The adaptation effect was determined from the PSE in the psychometric function for all angles between adapt location and test location to provide a spatial map of the spread of the FAE. Results show non-retinotopic adaptation – an FAE of similar strength at all locations -- following adaptation to a single face stimulus. Nevertheless, simultaneous adaptation to both the face and the anti-face, spaced at opposite locations across the fixation, did not cancel, producing instead two separate regions of opposite aftereffects. With the smaller separation of the 4 equally spaced stimuli, faces alternating with anti-faces, FAE was greatly reduced.


4:00 228 Grouping determines object-based selection in human inferior intra-parietal sulcus
Yaoda Xu1 (yaoda.xu@yale.edu), Marvin Chun1; 1Yale University
Visual objects are important units in visual processing, and yet how the brain determines what constitutes as a discrete visual object remains largely unknown. Previously, in a visual short-term memory (VSTM) study, we showed that, regardless of object complexity, fMRI response in the inferior intra-parietal sulcus (IPS) increased as the number of objects at different spatial locations increased (up to about four objects) (Xu & Chun, 2006, Nature). This finding indicates that the inferior IPS may participate in visual object individuation and selection, and its activation may reflect the number of discrete visual objects represented in the brain. In the current VSTM study, we asked: If object locations are fixed, but the grouping between objects differs, would that change the number of discrete objects represented in the brain? Our display consisted of two filled black rectangles on either side of the central fixation, similar to Egly, Driver and Rafal (1994, JEP:G). Observers encoded in VSTM two gray shapes appeared on the black rectangles and decided whether a probe shape at test matched one of the remembered shapes. Although grouping by closure (rectangles) was irrelevant to the VSTM task and could be ignored, the inferior IPS activation was lower when the two shapes appeared on the two ends of the same rectangle than on the same end of both rectangles. With an increased VSTM load, this grouping by closure allowed more shape information to be retained in VSTM as reflected in behavioral VSTM performance and fMRI responses in the superior IPS, which correlate with VSTM capacity. Interestingly, grouping did not modulate responses in the lateral occipital complex, an area involved in shape processing. Thus, grouping changes the number of discrete objects represented, and the inferior IPS may play a key role in determining object-based visual processing in the brain.

IPSとかparietalとfrontalの何にでも関係しているように見える部位の機能をパーツに分けて理解していくのはますます重要になるだろう。


2:45 439 Dynamic modulation of direction selectivity by task demands in prefrontal cortex
Cory Hussar1 (Cory_Hussar@urmc.rochester.edu), Tatiana Pasternak1; 1Dept. of
Neurobiology & Anatomy and Center for Visual Science
Responses of neurons in the prefrontal cortex (PFC) have been shown to represent behaviorally relevant sensory information. Recent work revealed that PFC neurons exhibit direction selective (DS) responses to visual motion used in a delayed match-to-sample (DMTS) task in which the monkeys compared two directions of motion separated by a brief delay (Zaksas & Pasternak, J. Neurosci, 2006). We asked whether this directionality is preserved when the monkeys were asked to ignore stimulus direction. We compared responses of PFC neurons to identical visual motion stimuli presented during three different tasks. In separate blocks of trials the animals discriminated the direction or the speed of motion, or were rewarded for passively maintaining fixation. We found that DS responses were drastically attenuated under both conditions in which stimulus direction was irrelevant and that the nature of this attenuation was task dependent. When the monkey discriminated stimulus speed and ignored its direction, DS activity was strongly attenuated early in the response and emerged about 100ms later than during the direction discrimination task. This early reduction of DS activity resulted not from a decrease in the response to the preferred direction, but from an increase in the response to the anti-preferred direction, suggesting an active release from inhibition characteristic of direction selectivity in visual cortical neurons. In contrast, during the passive fixation task, reduction in DS activity resulted largely from the decrease in the response to the preferred direction, which unlike the more transient loss in DS during the speed task, persisted throughout the response. Our results demonstrate the existence of a dynamic gating mechanism by which PFC neurons can modulate direction selectivity characteristic of visual cortical neurons.

途中からはいって聞いてて、これがprefrontalで記録してるって気づいて驚いた。


2:30 444 Non-retinotopic crowding
Patrick Cavanagh1,2 (patrick@wjh.harvard.edu), Alex O. Holcombe3; 1Harvard University, 2Université de Paris 5, 3University of Sydney

Sixteen radial arms are presented, each with a test letter at the third location from the center, flanked by 2 distractors on each side. The display counterphases targets and distractors in the following way: in one frame, every even arm (2nd, 4th, ..) shows only the target without distractors, while the odd arms (1st, 3rd, ..) show only the distractors without the targets; in the other frame, odd arms show distractors while even arms show the targets. When attending to any one location, a flickering test is seen, surrounded by flickering distractors. However, when a guide (a faint radial sector) moves from arm to arm in phase with the alternation, it contains only the test letter and no flankers. Targets are presented either in normal or mirror reversed orientation. Subjects fixate the center of the array and report the orientation of the test. When attention is directed to one fixed location, there is substantial interference from the flanking letters. However, when following the guide, crowding is much reduced, suggesting that crowding is specific to the flankers, if any, that move with the target and not to the letters that surround each target locally in retinotopic coordinates.

またもや、attentionをぐるぐるまわすパラダイムで、新しいことをひとこといっている。あの手法は予想以上にfruitfulだ。patrickにとってはcrowdingはattentionとbindingの問題だと考えている。こういう実験で示していくのはpromisingだ。

4:00 455 Entasis: architectural illusion compensation, aesthetic preference or engineering necessity?
Peter Thompson1 (p.thompson@psych.york.ac.uk), Georgia Papadopoulou1, Eleni Vassilou1; 1Department of Psychology, University of York, York, UK.

A typical characteristic of columns in Doric temples is entasis; a slight convexity in the body of the column. Often, particularly in guide-books, it is suggested that entasis compensates for an illusion of concavity in columns with parallel sides. Architectural experts generally agree: Entasis is the“swelling given to a column in the middle parts of the shaft for the purpose of correcting a disagreeable optical illusion, which is found to cause their outlines to seem concave instead of straight” - Penrose (1888). Nuttgens (1999) writes ‘Most Greek buildings of this golden period use entasis, the device whereby columns are given a slight swelling ....to counteract a tendency of the eye to see them as curving inwards from either side...’ We investigated whether any such illusion exists in a series of experiments in which we vary the degree of entasis from a negative value (columns waisted in the middle) through parallel sides to positive values (columns bulging in the middle). Our experiments presented 7 stimuli; 3 convex, 3 concave and 1 with truly parallel columns in a constant stimuli paradigm. Each stimulus was presented 30 times (to 12+ subjects) in pseudo-random order and results plotted as a psychometric function from which the PSE where columns appear neither convex nor concave determined. Several experiments, with more or less realistic column stimuli, failed to find any evidence to support any illusion-compensation theory. Secondly, we have explored the possibility that entasis was employed for purely aesthetic reasons. 5 computer-generated temples were judged for their aesthetic preference by 30 subjects. The temples differed in the application of entasis on the columns: 2 had negative entasis, 2 positive and one had parallel-sided columns. The results showed positive entasis as the least preferred aesthetically. Finally, we shall present some evidence supporting an engineering role for entasis.

entertainingなだけじゃなくて、今回は内容も印象に残った。一時期自分がアートとイリュージョンについて、考えて批評を読んでいたときは、vision scientistとして納得のいかない説明が多かった。そういう状況が文化についての学問ではよくあることだろう。そういうどうでもいい学問で、難しい文章でごまかすことを職業にしている人たちには腹が立つ。

D1 462 Stabilizing bistable visual patterns through interocular suppression, crowding, and inattention
Sheng He1 (sheng@umn.edu), Yi Jiang1, Xiangchuan Chen1; 1University of Minnesota
Alternating perception of bistable visual patterns can be slowed, and even halted, if the visual stimuli are presented intermittently and periodically removed from view. Here we demonstrate the similar stabilization effect of a number of different bistable visual patterns by interrupting the conscious experience of the bistable stimuli through interocular suppression, crowding, or removal of attention. Specifically, we found that: (1) Binocular rivalry: the perceptual switch of binocular rivalry could be dramatically slowed down by intermittently adding distractors around the rivalry stimulus and by intermittently engaging observers attention on a demanding task away from the rivalry stimulus. This observation is consistent with recent discoveries of attentional modulation of rivalry speeds. (2) Bistable rotating sphere: the perceived directional switch of a bistable rotating sphere defined by structure from motion patterns could be stabilized by intermittently rendering the bistable stimulus invisible through interocular suppression as well as through crowding. Inattention was also effective in stabilizing rotation. These observations imply that the neural stages of interocular suppression and crowding precede that of the structure from motion. (3) Bistable plaid: The perceptual alternation between two component gratings sliding across each other and a single coherently moving plaid could be stabilized with intermittent interocular suppression. Such an observation also suggests that interocular suppression disrupts information processing before the engagement of the neural mechanism responsible for component vs. pattern motion. Taken together, these observations indicate that interocular suppression, crowding, and the removal of attention are all nearly as effective as physical removal of bistable stimuli in perceptual stabilization. These observations also support the critical role of attention in facilitating perceptual alternations of bistable stimuli.

これは、自分でもまったくおなじ実験をして、同じデータを得ている。でも、ここまでしかsheng heでもいってないのかという印象だった。見えない間に、刺激に変化を与えて、stabilizationを壊す実験はしていないらしい。でも、1をいえば100ぐらいわかっている印象だった。

D22 483 Subjective area size influences time perception
Fuminori Ono1 (fuminori@med.juntendo.ac.jp), Jun-ichiro Kawahara2; 1Juntendo
University, 2National Institute of Advanced Industrial Science and Technology
The perceived duration of events is affected by non-temporal attributes of stimuli, such as the number of components, size, or complexity. Presenting an observer with more stimuli tends to result in an overestimated duration. Although previous studies have suggested that time perception is influenced by the physical attributes of stimuli, it is not known whether time perception is influenced by differences in the subjective appearance of physically identical stimuli. Specifically, we examined whether the subjective area size of the critical object in an Ebbinghaus illusion influences its time perception. We measured the perception of time spent looking at visual objects whose perceived area size was altered by the Ebbinghaus illusion, in which a central circle surrounded by large inducers appears to be smaller than a central circle of the same size surrounded by small inducers. In the experimental trial, one of two types of surrounding circle (the subjectively large or small conditions) was randomly displayed for 1500 ms. The central circle appeared for either 150 or 350 ms. The participants estimated the duration of the central circle. The results showed that the perceived duration of the subjectively large condition was longer than that of the subjectively small condition, although the actual area size remained invariant. This suggests that later visual processing systems influence time perception, because the Ebbinghaus illusion is a prototype for size contrast illusions that affect cognitive judgment by introducing bias into the processing of information at a higher level of visual processing (Coren & Enns, 1993). In summary, these results are the first to show the effect of a size contrast illusion (the Ebbinghaus illusion) on processing the temporal characteristics of a stimulus. This indicates that the time perception of visual events is influenced by higher-level representation of visual processing.

この実験はやるべきだとおもっていたけど、すごくきれいな結果が出ていた。theory of magnitudeを基にほかにも大きさのようなmagnitude的なものと時間の関係を調べたら面白い。

あと、日曜のポスターでの印象としては、binocular rivalryとかpsychophysicsでやることはもうすぐなくなるだろうという予測とは反対に、けっこうたくさんのひとがそれぞれ意味のある発見をしていることに驚いた。time perceptionは増えてはいるが、まだ爆発的な量になっていないから、それほど激しい競争なしで、おもしろい発見ができるフィールドだとおもった。rivalryはほとんど、アイデア的にはすでに考えたことのあるものばかりだった。

E74 623 Is residual vision in monkeys with unilateral lesion in the primary visual cortex like normal, near-threshold vision?
Masatoshi Yoshida1,2 (myoshi@nips.ac.jp), Kana Takaura1,2, Tadashi Isa1,2,3;
1Dept Developmental Physiol, Natl Inst Physiol Sci, Okazaki, Japan, 2School of
Life Science, The Grad Univ for Adv Stud, Hayama, Japan, 3CREST, JST,
Kawaguchi, Japan
Some of the patients with damages in the primary visual cortex (V1) retain their ability to localize visual targets in their affected hemifield (‘blindsight’or ‘residual vision’). One of the controversies about blindsight is whether it is a form of near-threshold vision or qualitatively different from normal vision. However, the extent of damage of the patients was not necessarily complete and restricted within V1, thus leaving the question unsolved. Here we used monkeys with unilateral lesion in V1 as an animal model of blindsight and examined whether the monkeys’ residual vision is a kind of attenuated vision that can be mimicked by presenting nearthreshold stimuli to their intact visual hemifield. After the unilateral removal of V1, two monkeys were tested with a visually guided saccade task with a target in one of five possible directions either in the intact or affected hemifield. The monkeys correctly localized the saccadic targets presented in the affected hemifield (70-95 % correct), while the accuracy of the saccade was worse than that of the intact hemifield, as quantified by the variance and the systematic errors in the saccadic end points. Then the monkeys were tested with near-threshold stimuli in the intact hemifield. Near-threshold stimuli were constructed either (1) by reducing luminance contrast or (2) by increasing spatial ambiguity. In both near-threshold conditions (70-90 % correct), the variance and the systematic errors in the saccadic end points were far smaller than that of the affected hemifield. The saccadic reaction time in the near-threshold condition was longer than that of the affected hemifield. These results suggest that the behavioral effects of V1 lesion are not limited to vision, but V1 lesion affected visuomotor processing including the saccade control system and/or decision process. We conclude that the residual vision of monkeys with V1 lesion is not like normal, near-threshold vision.

サルのV1をとってblindsight状態でsaccadeがどうなったかって言う話。2秒まってからもサッカードできるっていうことに一番驚いた。じゃあ、見えてるんじゃないかっておもうところだけど、coweyタスクでyes/noにするとnoと答えるということからたぶん見えてないだろう。クリストフの3秒待ち仮説は、この実験結果と合致していない。そこのところが考えさせられる。あと、日本でちゃんと意識の研究をしている人をみると、やや嬉しい。

E83 632 Personifying Inanimate Objects in Synaesthesia
Jonathan Carriere1 (jcarrier@watarts.uwaterloo.ca), Kelly Malcolmson1, Meghan
Eller1, Donna Kwan1, Michael Reynolds2, Daniel Smilek1; 1University of Waterloo, 2Trent University
We report a case study of an individual (TE) for whom inanimate objects, such as letters, numbers, simple shapes, and even furniture, are experienced as having richly detailed, personalities. TE reports that her objectpersonality pairings have been there for as long as she can remember, are stable over time, occur independent of her intentions, and that this is true even for novel objects. In these respects, her experiences denote synaesthesia. We show that TE’s object-personality pairings are indeed consistent over time; she correctly recognized 91% of the personality attributes for familiar objects (3.4 SD greater than the control mean of 47%), and 80% of the attributes for novel objects (2.3 SD greater than the control mean of 57%), when presented with a selection of attributes previously provided for the same or other objects. A qualitative analysis of TE’s personality descriptions revealed her personifications are extremely detailed and multidimensional, with familiar and novel objects differing in specific ways – familiar objects having more social characteristics than novel objects in particular. We also show that TE’s visual attention can be biased by the emotional associations she has with personalities elicited by letters and numbers. In a free viewing task the valence of TE’s object-personality associations had predicted effects on object fixation tendencies. On average, TE fixated negative objects less often than positive objects. She also demonstrated attentional capture by negative objects, fixating negative objects longer than positive objects. Controls showed no significant differences. These findings demonstrate that synaesthesia can involve complex personifications for inanimate objects, which can influence the degree of visual attention paid to those objects.

これおもしろそうだったけど、発表者を発見できず。気になったまま、おわり。

きょうは、午前のトークもできるだけ見ようっと。

金井





|

« ATOM & When Pathway | トップページ | VSS 終了 »

「学会・シンポジウム」カテゴリの記事

コメント

こんにちは。先日はVSSポスターに訪れてくださってどうもありがとうございました。今月のASSCは参加されますか? ASSCではCowey task中に記録した上丘ニューロン活動のデータを発表する予定です。ご都合がつくようでしたら、ぜひごらんください。

投稿: pooneil | 2007年6月 6日 (水) 16時41分

こんにちは。ASSCの発表、おもしろそうですね。ぼくは、今回はASSCにはいけなそうなので、残念ながら見に行けません。

いまだに、このまえVSSで教えていただいた、2秒まってもblindsightが起きるということに驚いています。

また、どこかで会えるといいですね。

投稿: 金井 | 2007年6月 8日 (金) 15時14分

pooneil さん、
私はASSC11行きますが、
2日目と3日目の午前は抜けなければならないので、
残念ながら pooneli さんのポスターは見れません。

abstract が面白そうなので、
もしよければ後で見せてください。

2日目のposterは、Melanie Wilke と Alex Maier も面白そうなポスターあるんですよね。

残念。

投稿: 土谷 | 2007年6月13日 (水) 13時38分

ここに土谷さんがコメントされていたことにASSC11から帰ってきたいま気付きました。ASSC11ではぶじ土谷さんとお会いして話すことが出来ました。こうやってブログが縁で繋がってゆくおもしろさを体験できたことに感謝してます。ではまた。

投稿: pooneil | 2007年6月28日 (木) 10時33分

コメントを書く



(ウェブ上には掲載しません)


コメントは記事投稿者が公開するまで表示されません。



トラックバック

この記事のトラックバックURL:
http://app.f.cocolog-nifty.com/t/trackback/207912/6416909

この記事へのトラックバック一覧です: VSS 途中経過:

« ATOM & When Pathway | トップページ | VSS 終了 »