TIME SERIES ANALYSIS OF PAIN SENSATION THRESHOLDS OF THE SKIN

Time series of pain sensation thresholds are analyzed by the time series analysis technique. Mechanical stimulus, indentation of a needle on the skin, was used for pain stimulus. Subjects were healthy male and female humans. Stimulation was done for one minute, and time series of pain sensation thresholds were obtained. First, the trends of pain thresholds for 50 minutes were analyzed by regression analysis. First-order coefficients of first-order regression showed, on average, that pain thresholds tend to increase about 0.2 or 0.3% per minute. Second, frequency of the time series of pain thresholds was analyzed. Autocorrelation functions suggested that there exists periodicity in pain thresholds. To confirm this, alternative stimulation with a pin and cylindrical needle was done. Cross correlation showed that these were highly correlated, and it was considered that there are periodical factors in pain thresholds. The periods of the peak value of the power spectrum of pain threshold time series were obtained from fifty experiments. This showed that pain thresholds fluctuate periodically during about 10 minutes. As periodicity of pain thresholds was found, factors which produce this phenomenon were investigated. From the gate control theory, it was supposed that such fluctuations of pain threshold were caused by the inhibitory effects of touch or pressure sensation. To investigate these effects, pain thresholds and touch thresholds were measured simultaneously. Cross correlation and the coherence showed that there was a positive correlation between pain and touch sensation. This suggests that the periodicity is not caused by peripheral inhibition effects in the spinal cord but is caused by the higher central nervous system. In order to infer the conditions of the brain, several physiological parameters were measured. They were alpha rhythm・beta rhythm・theta rhythm from EEG, EMG, heart rate and skin temperature. Experiments were performed for 100 minutes. A two-dimensional autoregressive model was introduced to the analysis. Each physiological parameter was regarded as the input signal, and the pain threshold was regarded as the output of the system. For evaluation the relative power contribution was used, which is the ratio of the contribution of input signal to the spectrum of the output. For all physiological parameters, the relative power contribution was less than 50%. When focused on the main period of pain threshold which is about 10 minutes, the relative power contribution of theta rhythm had its peak at this period though values were low. From this result, as theta rhythm appears at the early stage of the sleep, there is a possibility that some periodicity of drowsiness causes the fluctuation of the pain threshold.