Federation of European Neuroscience Societies
Eur J Neurosci. 2018;48:3171–3185. wileyonlinelibrary.com/journal/ejn | 3171© 2018 Federation of European Neuroscience Societies and John Wiley & Sons Ltd
Received: 24 January 2018 | Revised: 24 July 2018 | Accepted: 27 July 2018 DOI: 10.1111/ejn.14121
R E V I E W A R T I C L E
Neuronal correlates of motion- defined shape perception in primate dorsal and ventral streams
Takashi Handa1,2 | Akichika Mikami1,3
Edited by Dr. Helen Barbas. Reviewed by Georgia Gregoriou and Arash Yazdanbaksh.
All peer review communications can be found with the online version of the article.
Abbreviations: ITC, inferior temporal cortex; KB, kinetic boundary; LGN, lateral geniculate nucleus; LOC, lateral occipital complex; lSTS, lower bank of the anterior superior temporal sulcus; MRI, magnetic resonance imaging; MT, middle temporal area; PPC, posterior parietal cortex; RF, receptive field; SFL, shape from luminance; SFM, shape from motion; STP, superior temporal polysensory area; uSTS, upper bank of the anterior superior temporal sulcus; V1, primary visual cortex.
1Department of Behavioral and Brain Sciences, Primate Research Institute, Kyoto University, Inuyama, Japan 2Department of Behavior and Brain Organization, Center of Advanced European Studies and Research (CAESAR), Bonn, Germany 3Faculty of Nursing and Rehabilitation, Chubu Gakuin University, Seki, Japan
Correspondence Takashi Handa, Department of Behavior and Brain Organization, Center of Advanced European Studies and Research (CAESAR), Bonn, Germany. Email: takashi.handa@caesar.de
Abstract Human and non- human primates can readily perceive the shape of objects using visual motion. Classically, shape, and motion are considered to be separately pro- cessed via ventral and dorsal cortical pathways, respectively. However, many lines of anatomical and physiological evidence have indicated that these two pathways are likely to be interconnected at some stage. For motion- defined shape perception, these two pathways should interact with each other because the ventral pathway must uti- lize motion, which the dorsal pathway processes, to extract shape signal. However, it is unknown how interactions between cortical pathways are involved in neural mech- anisms underlying motion- defined shape perception. We review evidence from psy- chophysical, lesion, neuroimaging and physiological research on motion- defined shape perception and then discuss the effects of behavioral demands on neural activ- ity in ventral and dorsal cortical areas. Further, we discuss functions of two candidate sets of levels: early and higher- order cortical areas. The extrastriate area V4 and middle temporal (MT) area, which are reciprocally connected, at the early level are plausible areas for extracting the shape and/or constituent parts of shape from motion cues because neural dynamics are different from those during luminance- defined shape perception. On the other hand, among other higher- order visual areas, the an- terior superior temporal sulcus likely contributes to the processing of cue- invariant shape recognition rather than cue- dependent shape processing. We suggest that shar- ing information about motion and shape between the early visual areas in the dorsal and ventral pathways is dependent on visual cues and behavioral requirements, indi- cating the interplay between the pathways.
K E Y W O R D S dorsal stream, functional interaction, shape perception, ventral stream, visual motion
www.wileyonlinelibrary.com/journal/ejn
http://orcid.org/0000-0003-3956-8077
mailto:takashi.handa@caesar.de
3172 | HANDA AND MIKAMI 1 | INTRODUCTION Among mammals, primates heavily rely on vision. The visual systems of most primates have adapted evolutionally to diur- nal activity. For diurnal primates, visual object recognition plays pivotal roles in the judgment of good foods, such as ripe fruits, and in appropriate action selection, such as catching prey or escaping from predators (Barton, 1996, 1998; Kay & Kirk, 2000). Accordingly, shape perception is a fundamen- tal step in the processing of object recognition. Various vi- sual features, including luminance, color, texture, depth, and motion, enable human and non- human primates to perceive the shape of an object. For instance, visual motion cues are critical in detecting animals that camouflage themselves with a similar color and texture to their surroundings while the animals are still. Once they have moved, it becomes easier for observers to recognize them (Curio, 1976; Eckert & Zeil, 2001; Julesz, 1971; Robinson, 1969) (Figure 1a). In humans, relative motion is the most efficient cue for object segmenta- tion from a visual scene (Nawrot, Shannon, & Rizzo, 1996).
How does the primate brain perform shape perception using such motion cues? A classical view of the primate vi- sual system is that shape and motion are processed through
distinct pathways. Visual information is first transmitted from the retina to the cerebral cortex not only through the lateral geniculate nucleus (LGN) in the thalamus (Leventhal, Rodieck, & Dreher, 1981; Perry, Oehler, & Cowey, 1984; Schiller & Logothetis, 1990) but also through the superior colliculus and inferior pulvinar thalamic nucleus (Berman & Wurtz, 2010; Lyon, Nassi, & Callaway, 2010). Visual transmission through the retino- geniculo pathway has been anatomically and physiologically classified into two parallel pathways. The first is called the parvocellular (or color- opponent) pathway, in which the small receptive fields (RFs) of cells exhibit red- green color- opponent re- sponse patterns and the cells convey sustained signals with spatially fine resolution. A small lesion in the LGN par- vocellular layer has been shown to impair the detection/ discrimination of color, texture, and fine patterns. The sec- ond pathway is called the magnocellular (or broad- band) pathway, in which cells have large RFs and convey achro- matic, low spatial resolution, and more transient signals. A small lesion in the LGN magnocellular layer has been shown to impair motion perception (Derrington & Lennie, 1984; Schiller & Logothetis, 1990; Schiller, Logothetis, & Charles, 1990; Shapley & Perry, 1986). Thus, the parvocel- lular and magnocellular layers are capable of sending sig- nals for processing shape/color and motion, respectively. In the cerebral cortex, two visual pathways originating in the primary visual cortex (V1) have also been characterized. The parvocellular and magnocellular pathways are func- tionally correlated to the ventral and dorsal cortical path- ways, which have been considered to compute non- spatial (shape and color) and spatial (motion and depth) visual features, respectively (Ungerleider & Mishkin, 1982; Van Essen & Gallant, 1994). Among early visual cortical areas, the extrastriate area V4 in the ventral pathway and the mid- dle temporal (MT) area in the dorsal pathway have been extensively profiled. The V4 is critical for shape and color vision (Pasupathy, 2015; Roe et al., 2012), whereas the MT area is dedicated for processing visual motion (Born & Bradley, 2005). Among higher visual cortical areas, the ventral pathway terminates in the inferior temporal cortex (ITC) (Connor, Brincat, & Pasupathy, 2007; Tanaka, 1996; Tompa & Sáry, 2010), whereas the dorsal pathway is di- vided into two side streams that are linked to the posterior parietal cortex (PPC) (Goodale & Milner, 1992; Maunsell & Van Essen, 1983) and anterior superior temporal sulcus (Boussaoud, Ungerleider, & Desimone, 1990) (Figure 2).
For motion- defined shape perception, some corti- cal areas must use motion to extract the boundary be- tween the object and the background or shape of the object. The ventral and dorsal pathways seem to not be wholly independent; rather, they potentially inter- act with each other. In the ventral and dorsal cortical areas, neural inputs originating from the parvocellular
F I G U R E 1 A schematic illustration of motion- defined shape perception. (a) Left: A butterfly camouflaged by its surrounding when still. Once it moves, the shape can be detected by the primate visual system. The white arrow indicates the direction of movement of the butterfly. The white dashed line contour indicates the shape of the butterfly. Right: Extended view around a circle in gray in the left panel. The boundary (white dashed line) is visible by the movement of dots (arrows) on the butterfly against the still dots background. (b) In the laboratory, some artificial motion- defined form stimuli have been used. Left: The kinetic boundary (KB), a visible oriented line (dashed line) at the boundary between the counter movements of dots. Right: Shape from motion (SFM). The relative motion between the inside and outside field of an object enables us to see the shape (circle). Gray arrows indicate the direction of the movement of dots
(a)
(b)
| 3173HANDA AND MIKAMI
and magnocellular pathways physiologically and ana- tomically merge (Maunsell, 1992; Nassi & Callaway, 2009). Parvocellular layer inactivation reduces, but not completely eliminates, visual responses in the V4. Magnocellular layer inactivation comparably reduces the firing rate of V4 neurons in response to an oscil- lating white bar by approximately 40%. Thus, both LGN pathways contribute to visual responses in V4 (Ferrera, Nealey, & Maunsell, 1992; Ferrera, Nealey, & Maunsell, 1994). Moreover, non- direction- selective V4 neurons become tuned to the direction of random dot movement after monkeys have adapted to a visual mo- tion stimulus (Tolias, Keliris, Smirnakis, & Logothetis, 2005). Although the visual responses of MT neurons strongly depend on magnocellular contribution, the responsiveness of a few MT neurons reduced follow- ing parvocellular layer inactivation (Maunsell, Nealey, & DePriest, 1990). Rabies virus tracing has provided further evidence of multisynaptic innervations, which are disynaptic connections linking the magnocellu- lar pathway to the V4 and disynaptic connections the linking parvocellular pathway to the MT (Nassi, Lyon, & Callaway,…
