NeuroDetect publishes machine learning article showing effectiveness of using cellular electrophysiology to predict genetic mutation class in epilepsy
NeuroDetect is proud to announce the publication of our machine learning journal article showing the effectiveness of using in vitro (laboratory) measures of cellular electrophysiology from microelectrode arrays (MEAs) to predict genetic mutation class in epilepsy. This is a significant step forward in developing bench to bedside models to diagnose and treat neurological disorders and neurodegenerative diseases. Our journal article can be found on the NeuroDetect software page.
To produce this novel paper, we were proud to collaborate with the BioFutures Internship Program organized by BioscienceLA. Gavin Kress, MS, Chief Technology Officer of NeuroDetect, is first author of the paper. Fion Chan, BS, recent graduate of Cal Poly Pomona, and Claudia Garcia, BS, recent graduate of University of California, Irvine, made significant contributions. Following is a description of our paper.
Background: Epilepsy is the fourth-most common neurological disorder, affecting an estimated 50 million patients globally. Nearly 40% of patients have uncontrolled seizures yet incur 80% of the cost. Anti-epileptic drugs commonly result in resistance and reversion to uncontrolled drug-resistant epilepsy and are often associated with significant adverse effects. This has led to a trial-and-error system in which physicians spend months to years attempting to identify the optimal therapeutic approach.
It is well-established that genomic data alone can sometimes be predictive of effective therapeutic approach. Thus, to assess the predictive power of cellular electrophysiological data, machine learning strategies were implemented to predict a subject’s genetically defined class in an in silico model using brief recordings obtained from simulated neuronal networks.
Methods: A dynamic network of isogenic neurons was modeled in silico for one-second for 228 dynamically modeled patients falling into one of three categories: gain-of-function (GoF), loss-of-function (LoF) and healthy. Data from previous studies investigating the electrophysiological and cellular properties of neurons in vitro were used to define the parameters genetic mutation class. Ninety-two electrophysiological features defining the nature and consistency of network connectivity, activity, waveform shape, and complexity were extracted for each patient network and t-tests were used for feature selection to test five different machine learning algorithms.
Results: Our results showed that several machine learning algorithms excelled in using electrophysiological data from isogenic neurons to accurately predict genetic class with a Gaussian Naïve Bayes Classifier predicting healthy, gain of function, and overall, with the best accuracy, area under the curve, and F1. The Gradient Boosting Decision Tree performed the best for loss-of-function models indicated by the same metrics.
READ FULL ARTICLE HERE