Evidence From Learning With The Whiskers

The rat whisker sensory system is well suited to the inquiry because individual whiskers on the snout, arranged in rows and arcs, project to contralateral cortical "barrel-columns" in a precise topographic manner.30 Are all the barrel-columns of the cortical map bound together as a functional unit during learning or can different areas participate independently in sensory learning? Rats deprived of all but one whisker were trained in the "gap-crossing" task:31 they learned to use their remaining intact whisker to locate a "goal platform" and to cross the intervening gap to reach a food reward. Training was carried out under red light, preventing the albino rats from using visual information. They learned to feel the opposite platform across gaps of up to 16 cm. Once a rat had learned to gap-cross reliably, the "trained" whisker was clipped and a "prosthetic" one attached to the stub of an "untrained" whisker, or to the stub of the previously trained whisker.

FIGURE 7.3 (See Color Figure 7.3 in color insert.) Topographically distributed "response maps" for four whiskers of the rat's snout (C1-4) as revealed with large 10 x 10 microelectrode arrays inserted in somatosensory cortex. The four stimulus sites are shown on the drawing of the rat. To the right, the recording site is shown relative to the barrel field of the left hemisphere. The boxed area indicates the boundaries of the electrode array. At the bottom, spatial distributions of the four whiskers' representations are illustrated in relation to the barrel arrangement (right side). The center of each activated zone, where spike counts were highest, is given by the warmest color.

FIGURE 7.3 (See Color Figure 7.3 in color insert.) Topographically distributed "response maps" for four whiskers of the rat's snout (C1-4) as revealed with large 10 x 10 microelectrode arrays inserted in somatosensory cortex. The four stimulus sites are shown on the drawing of the rat. To the right, the recording site is shown relative to the barrel field of the left hemisphere. The boxed area indicates the boundaries of the electrode array. At the bottom, spatial distributions of the four whiskers' representations are illustrated in relation to the barrel arrangement (right side). The center of each activated zone, where spike counts were highest, is given by the warmest color.

With the prosthetic whisker in place, they were immediately tested. The main result was that rats gap-crossed without delay only if the prosthetic whisker was reattached to the stub of the previously trained whisker. For any other whisker transposition, a period of relearning was required: indeed, the number of trials necessary to reacquire the task increased systematically as a function of the distance along rows and arcs between the trained and the prosthetic whisker. Rats showed no benefit from their previous training if the prosthetic whisker was attached two or more positions away from the site of the original whisker.

Learning with a single whisker might seem to be a special case, but additional experiments demonstrated that the topographic learning rule was the same for multi-whisker learning. When rats were trained with a set of four adjacent whiskers, the learned ability quickly transferred to any of the whiskers immediately surrounding the trained ones, but transferred less readily to non-adjacent whiskers. Moreover, the topographic rule held up across the midline: rats could rapidly transfer learning to whiskers symmetrically opposite the trained whiskers (e.g., left C3 to right C3), but required additional training before successfully transferring to non-symmetric opposite whiskers. The corpus callosum, which connects homotopic sites in the left and right barrel field cortex,32 could be the neural substrate for this transfer.

From large-scale electrophysiological recordings using 100-microelectrode arrays implanted in cortex (Figure 7.3), we found that the extent to which learning transferred across whisker positions was perfectly correlated with the degree of overlap between the representations of those whiskers in primary somatosensory "barrel" cortex.

Based on the behavioral evidence, it is difficult to hold that the cortical network that participated in learning and remembering the tactile task was uniformly or globally distributed; in that case, the rats would have utilized the prosthetic whisker to gap-cross without delay, even if it were attached far from its original site. Instead, the observed pattern points to a memory trace governed by the precise topography of sensory cortex. The neurophysiological evidence confirms the feasibility and simplicity of this explanation.

Was this article helpful?

0 0

Post a comment