A recent study published online on February16th in the Proceedings of the National Academy of Sciences demonstrated the fine spatial structure of neuronal receptive field (RF) in awake monkey secondary visual cortex (V2)，using electrophysiological recording and computational modeling. In addition, optical imaging was used to reveal the cortical distribution of cells with different RF properties. This work was mainly performed by LIU Lu and SHE Liang in Dr. POO Mu-ming’s laboratory at the Institute of Neuroscience, Chinese Academy of Sciences.

　　Visual perception depends on the processing of the visual images through multiple brain regions in our brain. A crucial step in understanding visual processing mechanisms is to characterize the neuronal RFs at each stage, i.e. to reveal how visual signal is encoded and processed by single neurons at that stage.

　　Compared to the primary visual cortex (V1), the RF properties of the neurons in the secondary visual cortex (V2) are much less understood. The main obstacle is that the neuronal properties are highly nonlinear in higher cortices. Given the large number of potentially relevant visual features, traditional methods using stimulus sets with parametric variation of particular visual features are not efficient for comprehensive RF characterization. POO’s laboratoryhas used an alternative approach, which is to fit the stimulus-response relationship of each neuron by a parametric model (Fig. B). The neuronal responses were probed with large ensembles of natural images (Fig. A), and the parameters of the models were obtained through a computational method. The resulting model can be used to predict the neuronal responses to other arbitrary stimuli. Such an approach imposes no prior assumption about the stimulus selectivity of the cell and is thus well suited for unbiased RF characterization.

　　By using this approach, POO’s group discovered fine spatial structure of the V2 RF from the models of V2 neurons (Fig. C-E). They found that V2 RFs could be classified as “V1-like” (Fig. C), “ultralong” (high aspect ratios, Fig. D), or “complex-shaped” (with multiple oriented components, Fig. E). The latter two types were not reported previously, and their RFs could be explained by convergence of V1 neurons (Fig. F, G). This study illustrates that computational approach is useful for comprehensive RF identification in higher visual cortices.

　　Since imaging of intrinsic signals from V2 has revealed three different stripes in V2, the group collaborated with Dr. LU Haidong’s laboratory to combine information from electrophysiological recording with that from optical imaging. They found that neurons with ultralong RFs were primarily localized within pale stripes, whereas neurons with complex-shaped RFs were more concentrated in thin stripes.

　　This work was mainly performed by LIU Lu, SHE Liang, under the supervision of Dr. POO Mu-ming. The optical imaging experiment was performed by CHEN Ming under the supervision of Dr. LU Haidong. This work was supported by grants from Ministry of Science and Technology (973 program, 2011CBA00400) and Chinese Academy of Sciences (Strategic Priority Research Program, XDB02020001).

　　Figure legend: (A) Random sequences of natural images were presented and single-unit recordings were made in either V1 or V2. L.S., lunate sulcus; dashed line, V1/V2 border. (B) Linear-nonlinear RF model. Each subunit is represented by a linear filter (Fi) depicting the spatial RF structure. S represents stimulus. The response (xi = S？Fi) of each linear filter is passed through a second-order polynomial fi (xi) = ai + bi xi + ci xi² before summation. n, number of subunits; i, the i^th subunit; R, neuronal response. (C) Example V2 cell with V1-like RF. (D) Example V2 cell with ultralong RF. (E) Example V2 cell with complex-shaped RF. (F) A convergence model for V2 cells with ultralong RF . Red and blue ellipse: on and off sub-regions. (G) A convergence model for V2 cells with complex-shaped RF.