Normalization in human somatosensory cortex

Abstract

Functional magnetic resonance imaging (fMRI) was used to measure activity in human somatosensory cortex and to test for cross-digit suppression. Subjects received stimulation (vibration of varying amplitudes) to the right thumb (target) with or without concurrent stimulation of the right middle finger (mask). Subjects were less sensitive to target stimulation (psychophysical detection thresholds were higher) when target and mask digits were stimulated concurrently compared with when the target was stimulated in isolation. fMRI voxels in a region of the left postcentral gyrus each responded when either digit was stimulated. A regression model (called a forward model) was used to separate the fMRI measurements from these voxels into two hypothetical channels, each of which responded selectively to only one of the two digits. For the channel tuned to the target digit, responses in the left postcentral gyrus increased with target stimulus amplitude but were suppressed by concurrent stimulation to the mask digit, evident as a shift in the gain of the response functions. For the channel tuned to the mask digit, a constant baseline response was evoked for all target amplitudes when the mask was absent and responses decreased with increasing target amplitude when the mask was concurrently presented. A computational model based on divisive normalization provided a good fit to the measurements for both mask-absent and target + mask stimulation. We conclude that the normalization model can explain cross-digit suppression in human somatosensory cortex, supporting the hypothesis that normalization is a canonical neural computation.

Keywords: fMRI; forward model; normalization; somatosensory cortex; suppression.

Fig. 1.

Experiment design. A: vibrotactile stimulation was delivered to the right thumb (target stimulus) and right middle finger (mask stimulus). B: the visual task was a 2-interval change detection task. In the first interval (500 ms), 4 differently colored disks were presented around fixation. This was followed by a 500-ms interstimulus interval (ISI) and a second interval (500 ms) in which 1 of the 4 disks changed to a new random color. The subject was instructed to indicate which disk had changed in the subsequent response interval (1,000 ms), using the left-hand fingers to press 1 of 4 buttons.

Fig. 2.

Regions of interest (ROIs). Automated parcellation of human cortical gyri and sulci provided 4 ROIs: left and right hemisphere pre- and postcentral gyrus.

Fig. 3.

Activation and weight maps. A: activation map: stimulus-evoked activation shown on a flattened representation (flat map) of the left hemisphere central sulcus for a typical example subject. Dark gray, sulci; light gray, gyri; colors, activation of each voxel in the postcentral gyrus ROI, quantified as the fraction of the variance in the original time course that was accounted for by the regression model (r²). Dashed line indicates border between the post- and precentral gyri. B: weight map. Colors, weights assigned by the forward model to each voxel, rescaled to range from 0 to 1, for voxels with response reliability (from A) exceeding r² > 0.1.

Fig. 4.

Psychophysics. Subjects reported the stimulus interval containing vibrotactile stimulation to the thumb (target) as a function of target amplitude, with and without concurrent stimulation of the middle finger (mask). Filled symbols, target detection without mask; open symbols, target detection with mask. Error bars, SE across 5 subjects. Curves are maximum likelihood fits using Weibull functions.

Fig. 5.

Cross-digit suppression in human somatosensory cortex. A and B: channel responses from left hemisphere postcentral gyrus. C and D: right hemisphere postcentral gyrus. A and C: channel selective for thumb (target) stimulation. B and D: channel selective for middle finger (mask) stimulation. Filled symbols, mask absent; open symbols, target + mask. Error bars, SE across 6 subjects. Dashed and solid curves, normalization model.

Fig. 6.

Mean response amplitudes. A: mean response amplitudes, averaged across responsive voxels in the left hemisphere postcentral gyrus ROI. B: right hemisphere postcentral gyrus. Filled symbols, mask absent; open symbols, target + mask. Error bars, SE across 6 subjects. Dashed and solid curves, normalization model with cross-digit suppression.