Differences in sensitivity to neural timing among cortical areas

Abstract

The basic circuitry of auditory, visual, somatosensory and other cortical areas is highly stereotyped (Douglas and Martin, 2004). However, it remains unclear whether this anatomical stereotypy implies functional homogeneity, or whether instead different cortical areas are specialized to process the diverse sensory inputs they receive. Here we have used a two alternative forced choice task to assess modality-specific differences in the ability of rats to exploit precise neuronal timing. We delivered pairs of electrical pulses directly to different areas of cortex to determine the minimum timing differences subjects could detect. By stimulating the cortex directly, we isolated differences due to cortical circuitry rather than to sensory transduction and subcortical processing. Surprisingly, the minimum detectable timing differences varied over more than an order of magnitude, ranging from 1 ms in barrel cortex to 15 ms in visual cortex. Furthermore, these modality-specific differences depended upon sensory experience: although animals subjected to whisker clipping initially showed an impaired ability to exploit fine timing in barrel cortical stimulation, behavioral training partially rescued this deficit. Our results suggest that different cortical areas are adapted to the specific structure of the input signals they process, and that precise spike timing may play a more important role for some cortical areas than for others.

Figure 1.

A, Experimental paradigm. Rats were implanted with two electrodes placed 1.1 mm apart, and trained to discriminate different cortical stimulation patterns. Stimuli consisted of trains of 5 pulses delivered either simultaneously (AB) or sequentially (A-ISI-B) through the two intracortical electrodes, stimulation frequency 50 Hz. Animals initiated trials by poking into a center port, which elicited a stimulus after a 50 ms delay. They were rewarded for selecting the correct reward port. B, Brain slices showing rostral (left) and caudal (right) electrode positions in visual and barrel cortices. Arrows point to the electrode positions. Cytochrome oxidase (CO) staining was used for barrel cortex. Far right panel shows CO staining of flattened barrel cortex. Arrows point to the electrode positions. C, Training history of a representative visually implanted rat. The performance on successive training sessions is plotted in chronological order. Number on top indicates ISI (ms) of the training session. Error bars indicate binomial proportion confidence interval for a 95% confidence level. Filled circles, Sessions in which the animal performed significantly above chance. Empty circles, Sessions in which performance was at chance level. D, Training history of a representative barrel cortex-implanted rat. E, Training history of a representative auditory-cortex-implanted rat (from Yang et al., 2008).

Figure 2.

Different cortical areas differ in their ability to detect fine timing. A, Cumulative histogram showing fraction of animals able to discriminate each ISI for auditory (N = 26), visual (N = 10) and barrel (N = 6) cortex stimulation. The barrel cortex is most sensitive to fine temporal differences, followed by auditory cortex, followed by visual cortex. Auditory data reanalyzed from reference (Yang et al., 2008). B, The mean best performances as a function of ISI for different groups of rats recapitulates the sequence barrel>auditory>visual. Data were fit using cumulative Weibull Function (Wichmann and Hill, 2001; see Materials and Methods). C, Minimum detectable ISI for different cortices. Significance at 5% level by Fisher exact test.

Figure 3.

Sensory deprivation impairs the ability of barrel cortex to exploit fine neural timing. A, Sensory deprivation paradigm. Facial whiskers were trimmed unilaterally every 24 h from P0 to P60. Stimulation electrodes were later implanted in the contralateral barrel cortex (experimental group) or ipsilateral barrel cortex (control group 1) and nondeprived barrel cortex (control group 2). Control groups 1 and 2 did not show any difference in performance (p = 0.6, two-way ANOVA) and, therefore, were grouped together for analysis. B, Training history of a representative sensory deprived rat. C, Cumulative histogram showing fraction of barrel cortex-implanted control and sensory deprived animals able to perform each timing task. D, The mean best performances of control and sensory deprived rats.

Figure 4.

Impairment in fine timing discrimination of sensory deprived animals was partially rescued by training. A, Training history on ISI = 100 ms task of a barrel cortex control animal (blue) and a barrel cortex sensory deprived animal (red). Each closed or open point indicates one training session. Data were fit with an exponential function (see Materials and Methods). B, Cumulative histogram showing comparison of learning rates on the ISI = 100 ms task for control (blue) and sensory deprived (red) animals. C, Median performance of barrel cortex control, sensory deprived, and visual cortex-implanted animals at 100 ms across days. In normal animals (as opposed to sensory deprived animals), extensive training led to very little increase in performance. As is shown, visual cortex-implanted animals showed no improvement even after 2 weeks of training.