Hippocampal slice cultures spontaneously develop chronic epilepsy several days after dissection and are used as an in vitro model of post-traumatic epilepsy. This work describes the development of a hybrid microfluidic-microelectrode array device that improves the throughput of chronic recordings in hippocampal slice cultures and facilitates antiepileptic drug discovery. Our technology allows miniaturization of large and expensive multiple-slice electrophysiology systems to a single scalable chip. We used this epilepsy-on-a-chip device to carry out a screen of Receptor Tyrosine Kinases (RTKs) inhibitors and discovered two novel antiepileptic compounds. These ‘hits’ represent a promising first step in developing new antiepileptic drugs.
We present a novel Granger causality estimator based on recurrent neural networks (RNNs) to deal with the multivariate brain connectivity detection problem. The varying-length propagation delay between brain signals poses a challenge for classical approaches using fixed time lags. Our method takes time series signals with arbitrary length of transmission time lags and learns the information flow from the data with RNNs. The proposed method works well on varying-length time lags and even on very long transmission delays in electroencephalography (iEEG) signals, and is promising to serve as a robust brain connectivity analysis tool in clinical applications.
Seizure prediction is currently a major concern in the epilepsy research community. In this work, we have proposed a new strategy to achieve accurate seizure forecasting by combining effective connectivity measures and artificial intelligence techniques. Results show performance improvement compared to previous studies, achieving average sensitivity of 84.82% and time in warning of 0.1.
Our work emphasizes the understanding of clinical considerations and proper translation of these clinical considerations into data-analytic modeling assumptions. Several design choices during preprocessing and postprocessing are investigated for their effect on seizure prediction accuracy.
The epileptogenic zone (EZ) is a brain region responsible for seizure genesis. This study describes a novel EEG source imaging (ESI) method to estimate the EZ which uses cross frequency coupled potential signals (SCFC) derived from scalp EEG. Results were validated using 1) known surgical resections for Engel I-IV patients, and 2) through forward modelling with noise simulation. The SCFC demonstrated significant advantages over “raw” scalp EEG.
High-frequency oscillations (HFOs) are considered to be a good marker of the tissues which have to be removed. We investigated several methods to flatten (i.e. whiten) the spectrum to improve HFO detectability. Our method, the H0 z-score, provides an optimal framework for representing and detecting HFOs, independent of a baseline and a priori frequency bands.
We proposed a new epileptic seizure prediction method utilizing heart rate variability (HRV) analysis. It monitors time-frequency-domain HRV features for predicting seizures by using multivariate statistical process control (MSPC). The application results to clinical data produced accurate predictions (91%) for epileptic seizures and there were few false-positives (0.7 times/hour). The possibility of realizing a HRV-based epileptic seizure prediction system was shown.
With the advent of new recording techniques, fast electrical activity (>80 Hz) has become a new focus in brain research. We examined the relationship between the spatiotemporal coherence patterns of slow and fast electrical oscillations in the brains of patients with extratemporal lobe epilepsy.
we investigated the neurovascular/metabolic coupling in the epileptogenic cortices of rats with chronic focal epilepsy during ictal periods using a metabolically coupled balloon model, in comparison to normal rats undergoing forepaw stimulation. The results suggest that epileptogenic cortices have significantly larger hyperemic responses and a higher baseline of oxygen metabolism than normal cortices…
Advanced time-variant and non-linear approaches can be beneficially used for analysis of biomedical signals and are able to gain new insights into specific signal characteristics like short term patterns in heart rate variability…