Flow stimuli reveal ecologically appropriate responses in mouse visual cortex

Abstract

Assessments of the mouse visual system based on spatial-frequency analysis imply that its visual capacity is low, with few neurons responding to spatial frequencies greater than 0.5 cycles per degree. However, visually mediated behaviors, such as prey capture, suggest that the mouse visual system is more precise. We introduce a stimulus class-visual flow patterns-that is more like what the mouse would encounter in the natural world than are sine-wave gratings but is more tractable for analysis than are natural images. We used 128-site silicon microelectrodes to measure the simultaneous responses of single neurons in the primary visual cortex (V1) of alert mice. While holding temporal-frequency content fixed, we explored a class of drifting patterns of black or white dots that have energy only at higher spatial frequencies. These flow stimuli evoke strong visually mediated responses well beyond those predicted by spatial-frequency analysis. Flow responses predominate in higher spatial-frequency ranges (0.15-1.6 cycles per degree), many are orientation or direction selective, and flow responses of many neurons depend strongly on sign of contrast. Many cells exhibit distributed responses across our stimulus ensemble. Together, these results challenge conventional linear approaches to visual processing and expand our understanding of the mouse's visual capacity to behaviorally relevant ranges.

Keywords: flow movie; mouse; receptive field; spatial frequency; visual cortex.

Fig. 1.

Introducing flow stimuli. (A) Ecological motivation. Working with a single frame (A, 1–3), a grassy patch is modified to emphasize salient higher contrasts. Our abstraction (the flow field in A, 4) approximates this with a binary pattern of random, oriented dotted segments. (B) We generalize to flow fields consisting of dotted segments of different lengths, emphasizing two geometries [oriented (three or four aligned dots) or nonoriented (single dots)], two contrast polarities (positive or negative), different contrast magnitudes, and various sizes. The full flow stimulus is a movie of one such flow field, drifting across the screen with small random perturbations to suppress rigidity (Materials and Methods) (examples in SI Appendix, Movies S1 and S2). (C) Flow responses are inconsistent with classical filtering views of V1. Shown is a Gabor receptive field at 0.04 cpd superimposed onto the three-dot flow whose energy peaks at 0.24 cpd (example in B, Top Right), for comparison. (D) The 1D discrete Fourier transforms (single sided) of the flows used in our experiments (peaks at 0.15 cpd, 0.24 cpd, 0.7 cpd, 1.0 cpd, 1.25 cpd, and 1.6 cpd) have power well beyond 0.04 cpd (dashed curve), which is the spatial frequency previously reported as optimal for cells in mouse V1 (compare D, Inset). Adapted with permission from ref. . To compare stimuli, each spectrum is normalized by the power at the peak frequency (norm) (2D spectra in SI Appendix, Figs. S14 and S15).

Fig. 2.

Variety of responses in V1. (A–C) Tuning curves and PSTHs of three example cells in response to drifting gratings and flows at 0.04 cpd, 0.15 cpd, and 0.24 cpd in eight equally spaced directions of motion. Time axis in histograms encompasses an entire period of stimulus presentation (1.5 s). Insets in STAs show, at the same scale, stimuli that produced the most significant responses. (A) Cell responding to low-frequency gratings only. Bin size is 34 ms. (B) Cell responding preferentially to single-dot flows with negative contrast. Bin size is 83 ms. (C) Cell responding strongly to both oriented (three dots), positive flows and gratings (at both high and low spatial frequencies). Bin size is 46 ms. (D) Distribution of optimal spatial frequency in terms of proportion of cells significantly responding to at least one of the stimuli. In the group of experiments using the first set of stimuli (D, Left, 0.04–0.24 cpd, n = 357 cells, three animals), the majority of cells fired more strongly for stimuli at 0.15 cpd, followed closely by 0.04 cpd. For the second set of stimuli (D, Right, 0.04–1.6 cpd, n = 256 cells, three animals) there was an overwhelming preference for 0.04 cpd, although more than half the cells had optimal spatial frequency in the range 0.7–1.6 cpd. (E) Distribution of preferred stimulus among all cells. When low-frequency gratings (0.04 cpd) are included among the stimuli (E, Left), the majority of cells respond equally well to both classes (“Multi”), followed by only flows and only gratings; 29% of the cells were not significantly responsive (“N.S.”) to any of the stimuli displayed (n = 1,026 cells; 10 experiments, six animals). When we do not include low-frequency gratings (grat), thereby limiting the comparison with flows and gratings with similar spatial frequencies, there is a significant preference for flows only and for both over gratings only. Comparison of E, Left and Right reveals that ∼20% of cells preferred low-frequency gratings. When we recompute stimulus preference considering only stimuli with comparable spatial frequencies, most cells that preferred low-frequency gratings now either prefer none of the high-frequency stimuli or significantly prefer flows over high-frequency gratings, given that the fraction that prefers both remains essentially constant in the two scenarios. *P = 0.025; **P < 0.001. Error bars represent SEM. (F) Distribution of preferred stimulus among well-tuned cells (i.e., those with OSI > 0.5 or DSI > 0.5), n = 295 cells (Left) and 241 cells (Right); 8 experiments, four animals. Here, note that most of the cells responding to orientation and/or direction will fire more strongly to low-frequency gratings; F, Right reveals, however, that the fraction of cells well tuned to flows is just as large. And, similarly to E, many of the well-tuned cells preferring 0.04 cpd gratings prefer flows to gratings of comparable spatial frequency. *P < 0.05; **P < 0.001; ***P < 0.0001. Error bars represent SEM.

Fig. 3.

Cells remain highly selective at higher spatial frequencies. (A) Example of cell exhibiting a stronger response to oriented, negative flows at 0.7 cpd and 1.0 cpd compared with gratings at various spatial frequencies. Bin size is 47 ms. (B) Overall proportion of well-tuned cells among cells significantly responsive to each spatial frequency (Mann–Whitney U test, P < 0.05), irrespective of stimulus class. Sample sizes: 0.04 cpd (n = 508), 10 experiments, six animals; 0.15 cpd (n = 385), 0.24 cpd (n = 365), 5 experiments, three animals; and 0.7 cpd (n = 214), 1.0 cpd (n = 214), 1.25 cpd (n = 186), 1.6 cpd (n = 173), 5 experiments, three animals.

Fig. 4.

Cells in different layers have distinct selectivity toward different stimulus classes, as measured by a SSI (main text). ***P < 0.001; **P < 0.005; *P < 0.05. Error bars represent SEM.

Fig. 5.

Preference over flow stimuli variations. Percentages refer to the population of cells that had significant response to at least one flow variation. (A and B) All cells: n = 667, 10 experiments, six animals. Well-tuned cells: n = 187, 8 experiments, four animals. Error bars represent SEM (*P < 0.001). (A) Flow geometry preference. Among all cells responding significantly to flows, most showed no significant prefernce for either type. Among well-tuned cells, oriented flows were preferred over nonoriented flows. (B) Flow contrast polarity preference. Among all cells significantly responding to flows, positive polarity was preferred. The population of well-tuned cells showed no overall preference for contrast polarity. n.s., nonsignificant.