Clustering of neuronal response profiles in graded thermal stimulation

Functional data analysis has received significant attentions from various fields. In the present study, we performed the functional data analysis to characterize the responses of 147 spinal cord dorsal horn neurons to graded thermal stimuli, which vary from 37 to 51°C with increment of 2°C. We proposed a screening algorithm based on a multiple testing procedure controlling the false discovery rate to remove the neurons that behave constantly over graded thermal stimuli. For the remaining neurons, we estimated the B-spline functions that represent their response patterns to thermal stimuli. A clustering algorithm was applied to the coefficient estimates of the B-spline functions to elicit nature groupings of the neurons that have the similar response patterns to graded thermal stimuli. Four groups of neurons with distinct patterns were identified.