| Summary: | Given the nonlinearities of the the neural circuitry's elements, we would expect corticalcircuits to respond nonlinearly when activated. Surprisingly, when two points in the motorcortex are activated simultaneously, the EMG responses are the linear sum of the responsesevoked by each of the points activated separately. Additionally, the corticospinal transferfunction is close to linear, implying that the synaptic interactions in motor cortex must beeffectively linear. To account for this, here we develop a model of motor cortex composedof multiple interconnected points, each comprised of reciprocally connected excitatory andinhibitory neurons. We show how nonlinearities in neuronal transfer functions areeschewed by strong synaptic interactions within each point. Consequently, thesimultaneous activation of multiple points results in a linear summation of their respectiveoutputs. We also consider the effects of reduction of inhibition at a cortical point when oneor more surrounding points are active. The network response in this condition is linear overan approximately two to three fold decrease of inhibitory feedback strength. This resultsupports the idea that focal disinhibition allows linear coupling of motor cortical points togenerate movement related muscle activation patterns; albeit with a limitation on gaincontrol. The model also explains why neural activity does not spread as far out as the axonalconnectivity allows, whilst also explaining why distant cortical points can be, nonetheless,functionally coupled by focal disinhibition. Finally, we discuss the advantages that linearinteractions at the cortical level afford to motor command synthesis.
|
|---|
