Mapping retinotopic structure in mouse visual cortex with optical imaging

Abstract

We have used optical imaging of intrinsic signals to visualize the retinotopic organization of mouse visual cortex. The functionally determined position, size, and shape of area 17 corresponded precisely to the location of this area as seen in stained cortical sections. The retinotopic map, which was confirmed with electrophysiological recordings, exhibited very low inter-animal variability, thus allowing averaging of maps across animals. Patches of activity in area 17 were often encircled by regions in which the intrinsic signal dropped below baseline, suggesting the presence of strong surround inhibition. Single-unit recordings revealed that this decrease of the intrinsic signal indeed correlated with a drop of neuronal firing rate below baseline. The averaged maps also greatly facilitated the identification of extrastriate visual activity, pointing to at least four extrastriate visual areas in the mouse. We conclude that optical imaging is ideally suited to visualize retinotopic maps in mice, thus making this a powerful technique for the analysis of map structure in transgenic animals.

Fig. 1.

Imaging cortical retinotopy in mice.A, For retinotopic stimulation, we used square-shaped windows of gratings (25° side lengths) at adjacent positions within the visual field. The color code representing stimulus position was used to generate the color-coded retinotopic maps. Thewhite line indicates the vertical midline.B, Schematic of the imaged cortical region as indicated by the red window, which contains a rough outline of area 17. C, Blank corrected single-condition maps from one animal, with each map corresponding to a stimulus inA, according to the indices. All maps are scaled and clipped to the same absolute values. The cortical blood vessel pattern imaged through the translucent skull is shown at the bottom left. The map next to the blood vessel image displays the difference between images taken during two independent blank screen presentations. D, Reproducibility of the imaged maps from a different animal. Both sets corresponding to stimulib2–7 were averaged across 24 repetitions per stimulus imaged during subsequent blocks of data acquisition. Theelongated white region visible in the bottom right of some of the maps is most likely an artifact caused by the venous sinus. The blood vessel pattern is shown at theleft. Note that cortical blood vessels are clearly visible in the anterior part of the imaged region, whereas they appear very blurred in the posterior part, because of a different structure of the bone above this region. Despite this, activity maps could be readily imaged in this region as well. E, Anatomical verification of imaging in the primary visual cortex. The superficial blood vessel pattern (top panels) was used to align SMI-32 staining reflecting the anatomical position of area 17 (bottom left panel), with the maximum intensity projection of the intrinsic signal across all single-condition maps (bottom right panel). The red crosses were placed at the same positions in the staining pattern and the imaged map. F, Color-coded map of the overall retinotopic organization of area 17: the colorof each pixel corresponds to the stimulus position that elicited the strongest signal at this pixel. To mask out regions without cortical response, color saturation equals the maximum intensity projection of all single-condition maps. Scale bars, 1 mm.

Fig. 2.

Averaging maps across animals. A, Single-condition maps averaged across seven animals. Indices denote stimulus position according to Figure 1A.B, Color-coded retinotopic map based on the averaged single-condition maps. C, Spatial distribution of the SEM for each map. For better visibility, the SEM maps are scaled by a factor of 5 in comparison with the single-condition maps inA. The overall dark appearance of the maps indicates that the variability between animals is low. D, Histogram of the distances between receptive field positions (i.e., the maximal response) determined electrically and optically based on recordings from 22 neurons (black bars). The control distribution (white bars) was calculated from randomly assigning receptive field positions for 22 pairs and calculating the histogram of their distances. Note that 50% of the electrically measured receptive field positions coincide with the optically determined position. Scale bars, 1 mm.

Fig. 3.

Distribution of the CMF in area 17.A, Layout of the averaged retinotopic map in area 17. Values to the left denote elevation relative to the optic disk projection, and values at the bottom indicate azimuth relative to the vertical midline (which coincides with the midline between the two optic disk projections). The gridlines display isoelevation and isoazimuth lines for the contralateral eye. The approximate position of the horizontal meridian (Wagor et al., 1980;Dräger and Olsen, 1981) is indicated by the dashed line. The red (contralateral eye) andgreen (ipsilateral eye) crosses denote the averaged positions (n = 7 mice) of the centers of mass of the primary patches in area 17. The lengths of the crossing bars indicate 1 SEM in vertical and horizontal direction. Scale bar, 1 mm. B, C, Distribution of the CMF along the vertical (B) and horizontal (C) axes for each stimulus position (conventions as in A) coded by lightness.

Fig. 4.

Optical and electrical recording of lateral inhibition. A, In the top row, the track positions are marked in the color-coded retinotopic maps from three animals. The color code for stimulus position used for experiments with combined electrical and optical recordings is shown at the top right. The bottom part of this panel depicts two-dimensional plots of position tuning curves recorded electrically (dark gray background) and optically (light gray background) at the positions of the electrode tracks. Eachcolored circle displays the response to the corresponding stimulus position. The radius of each circle is proportional to the absolute value of the normalized firing rate or the position-tuning curve reconstructed from the optical signal at the track position. Red (blue) color indicates supra (sub) blank response. The circlecorresponding to the stimulus, which elicits maximal (minimal) response, is colored in yellow(turquoise); see also the legend at thebottom left. B, Histogram of the distances between the positions of the peaks of excitation (yellow bars, n = 22) and inhibition (turquoise bars, n = 17) determined electrically and optically. The control distribution (white bars) was constructed from randomized pairs of positions. Scale bars, 1 mm.

Fig. 5.

Retinotopic organization of extrastriate visual areas. A, High-pass-filtered, averaged single-condition maps (boxcar filter = 800 μm; n = 7). Thecolored arrows were placed at the same positions in all maps; they indicate distinct patches in extrastriate areas (yellow: area LM; green: area AL;red: area A; blue: area AM). Note that in many images several extrastriate patches are visible. The pronouncedwhite ring surrounding the dark patch in area 17 was introduced by the strong high-pass filter applied to these maps; it does not indicate the shape of the inhibitory surround.B, Color-coded retinotopic map using the high-pass-filtered single-condition maps shown in A. The color corresponds to the code for the average position projection. Color saturation is proportional to the scaled maximum intensity projection of the high-pass-filtered maps. Candidate extrastriate areas were outlined on the basis of the presence of separate patches in the single-condition maps in A. Scale bars, 1 mm.