Non-invasive detection of optical changes elicited by seizure activity using time-series analysis of light scattering images in a rat model of generalized seizure.

For the first time, we detected optical changes elicited by seizure activity in pentylenetetrazol (PTZ)-treated rats (n=6) versus saline controls (n=2) over a 30min recording session using a novel time-series analysis of scattering images obtained non-invasively with a real-time multispectral diffuse optical tomography (DOT) system. Spatio-temporal images of absorption and scattering coefficients were recovered from PTZ- and saline-treated rats' brains using a finite element-based DOT image reconstruction algorithm. After pulse artifacts were eliminated, an independent component (IC) analysis was conducted for blind-source separation of the optical signals. The retrieved ICs were compared with concurrently measured EEG signals, and the selected components were further refined using K-means clustering and spectrum analysis tools. The results revealed that changes in absorption and scattering coefficients emerge sooner than changes in the EEG signal and a low frequency peak signal of ∼0.3Hz in the spectra of light scattering images after PTZ injection. This low frequency caused by slow volume changes in CNS cells was not detected in control animals. Brain regions that we detected early changes in optical signals and activation maps were confirmed in an additional 3 PTZ-treated rats using the DOT system and concurrent EEG recordings obtained from multiple brain regions. Our results show that the analysis of scattered diffuse light is a sensitive and reliable modality for detecting changes in neural activity associated with generalized seizure and other CNS disorders with the additional benefit of providing access to physiological parameters that other modalities cannot access.