P 170 A potential signature for ongoing pain in mice

Abstract

Chronic pain is a major health care problem nowadays with no convincing treatment yet ( Apkarian et al., 2009 ). One challenge of research and treatment of pain is that it cannot be objectively quantified. Only indirect measurements like the Visual Analogue Scale for humans and responses to mechanical or thermal stimuli for animals are used. Electrophysiological changes in brain activity are a possible candidate for a direct measurement of pain and could be easily obtained by EEG recordings. However most of the studies with this scope are focused on evoked potentials of short lasting pain stimuli (ms to s) ( Zhang et al., 2012 ) and only few studies have been conducted on longer lasting pain ( Leblanc et al., 2014 ). Therefore we analyzed Local Field Potential (LFP) recordings in mice after injection of capsaicin in order to find a signature of ongoing pain. LFP was recorded in freely moving and head-restrained mice before, during and after ongoing pain induced by capsaicin injection into the hindpaw. As a control, NaCl was injected into the same hindpaw one day earlier. Implantation sites were the Primary Somatosensory Cortex - Hindlimb Region (S1HL), the Cingulate Cortex (ACC), the Ventro Posterolateral Thalamic Nucleus (VPL) and the Insular Cortex (Ins) both ipsi- and contralateral to the injection site respectively, as well as the Parietal Cortex and the Olfactory Bulb (left side only). For analysis the signal 15 min before injection of NaCl and immediately after injection (up to 5 min) was taken. To avoid muscle artefacts, phases in which the mouse did not move were extracted and concatenated until a 30 s window was reached. The mean power was then computed in six different frequency bands (1–4 Hz, 4–12 Hz,12–30 Hz, 30–80 Hz, 80–120 Hz, 120–160 Hz) using the MATLAB pwelch function (Hamming window 4s, 50% overlap). These values were normalized using the following formula (Power post Injection − Power pre Injection) /(Power post Injection + Power pre Injection) . A logistic regression model was fitted for each channel trying to predict whether capsaicin was administered based on the normalized power values across all frequencies. From these values a receiver operating characteristic (ROC) curve was computed and the area under the curve was taken as the probability of prediction whether capsaicin was injected. A p -value was computed using the wilcoxon signed rank test ( Table 1 ). To control for multiple testing the Benjamin-Hochberg procedure was used. We conclude that LFP recordings could be a measurement to infer whether an animal is experiencing acute pain and thus might be used as a signature of pain. However, a larger sample size is needed to determine the details of those changes in LFP activity and the most relevant frequency bands.