DISTRIBUTED ASPECTS OF THE RESPONSE TO SIPHON TOUCH IN APLYSIA - SPREAD OF STIMULUS INFORMATION AND CROSS-CORRELATION ANALYSIS

We examined two aspects of the response to siphon stimulation in an at tempt to test the hypothesis that the Aplysia CNS functions as a distr ibuted system. First, we estimated the number of central neurons that respond to a light touch to the siphon skin. We made voltage-sensitive dye recordings from the abdominal, pleural, pedal, and cerebral gangl ia. From these recordings we estimated that 220 abdominal neurons, 110 pleural neurons, and 650 pedal neurons were affected by the light tou ch. Thus, the information about this mild and localized stimulus is ve ry widely distributed within the Aplysia CNS. This result allows the p ossibility that the Aplysia CNS functions as a distributed system. If only a small number of neurons had responded to the touch, it would ha ve supported the conclusion that the gill-withdrawal reflex could be g enerated by a small, dedicated circuit. Second, we searched for correl ations between the spike times of the individual abdominal ganglion ne urons. Two time scales were examined: a millisecond time scale corresp onding to the duration of a fast synaptic potential and a seconds time scale corresponding to the duration of the gill-withdrawal movement. Neuron pairs with highly correlated spike activity on a millisecond ti me scale must be connected by (or have a common input that uses) relat ively powerful, fast, excitatory synapses. We expected that this kind of synaptic interaction would be relatively rare in nervous systems th at functioned in a distributed manner. Indeed, only 0.3% of the neuron pairs had correlation coefficients of 0.15 or greater. These correlat ions accounted for approximately 2% of the action potentials generated in response to siphon stimulation. Thus, large, fast excitatory synap tic interactions appear to be relatively unimportant. This result is c onsistent with the hypothesis that the abdominal ganglion functions as a distributed system. When the longer time scale was used for the cro ss-correlograms, a large fraction of the cell pairs had correlated act ivity because many neurons are activated by the stimulus. It was not p ossible to interpret the slow correlations in terms of actual synaptic interactions between individual neurons. Our results are consistent w ith the possibility that the abdominal ganglion functions in a distrib uted manner. However, this evaluation is indirect and thus only tentat ive conclusions can be drawn. Evidence from several sources suggests t hat the neuronal interactions for generating the Aplysia gill-withdraw al reflex are complex.