Activity-dependent cortical plasticity may contribute to organising the representation of sensory information in well-ordered maps. The rodent neocortical somatosensory map is exceptionally well delineated by the layer 4 barrel pattern, which is isomorphic to the layout of the whiskers on the rodent snout. These barrels can be visualised in living brain slices, allowing the neocortical circuits of specific regions of the sensory map to be investigated in detail. Activity-dependent plasticity can be induced by simply trimming whiskers.

In our experiments we trimmed whiskers belonging to rows A, B and C (sparing whiskers of rows D and E) for a period of over 10 days, beginning around a week after birth. The neuronal networks activated by the spared D row whiskers are particularly interesting to investigate since they are flanked by deprived cortex on one side (representing the trimmed C row whiskers) and spared cortex on the other (representing the spared E row whiskers). Large-scale changes in the representation of D row whiskers in layer 2/3 were investigated in vivo by voltage-sensitive dye (RH1691) imaging and whole-cell recordings targeted to specific barrel columns. Brief deflection of the D2 whisker in the urethane (1.5 g kg^-1)-anaesthetised rat evoked responses initiating in the D2 barrel column and subsequently spreading over the barrel field. In control animals the response preferentially spread towards the C row (n = 10), whereas in the deprived animals responses preferentially spread to the E row (n = 10). To investigate in detail how the neocortical circuits are altered, barrel columns representing rows C, D and E were identified in brain slices. Voltage-sensitive dye responses evoked by electrical stimulation of a D row barrel in vitro were also different between control (n = 19) and deprived (n = 20) animals. The early response was columnar in both groups but subsequently the signal spread asymmetrically within layer 2/3. In deprived animals (but not in control animals) the spread was preferentially towards the E row. This result suggests that changes in the layer 2/3 neuronal network might account for a major part of the observed plasticity. Thus the connectivity of the excitatory neuronal network of layer 2/3 was investigated by simultaneously recording postsynaptically from pyramidal neurons located in the C row and E row and checking for excitatory responses evoked by action potentials in individual pyramidal neurons located in the D row. The number of synaptic connections originating from D row pyramidal neurons was highest with C row postsynaptic neurons in control animals but with E row neurons in deprived animals.

This whisker trimming protocol thus induces a reversal of the preferred connectivity of excitatory neurons in the D row layer 2/3 barrel columns. After deprivation pyramidal neurons in layer 2/3 of D row preferentially synaptically connect with pyramids in the neighbouring spared E row barrel column rather than expanding into the deprived C row cortex.

All procedures accord with current national guidelines.