Statistical analysis of temperature-dependent neuronal activity

Abstract

A set of computer programs which collect and analyze data on the temperature dependence of neuronal activity is described. Data on the response to thermal stimultion of cultured neurons of the rat preoptic area were collected and displayed on-line and were stored on disk for off-line analysis. Statistical analysis of the data examined two alternate models of thermosensitivity: one that the relationship between neuronal activity and temperature was well fit by a linear relationship and the other, that it was well fit by semi-log regression. For each of the models, a neuron was considered to have temperature-dependent activity if two criteria were met: first, that the model was considered adequate by a lack-of-fit F-test and second, that the slope of the regression line differed from zero using a t-test. A statistical method was also used to test the hypothesis that the neuronal response to temperature displays a change point. This method did not identify any neurons whose thermal response relationship could be better fit by two intersecting regression lines than by one regression line.