You are here:
Semi-supervised Classification Algorithm using k-means Trajectory Clustering and Support Vector Machine to identify Neuroactive compounds from microelectrode array recordings of cortical neural networks
Citation:
Davila-Montero, S., T. Shafer, K. Wallace, J. Fink, AND A. Mason. Semi-supervised Classification Algorithm using k-means Trajectory Clustering and Support Vector Machine to identify Neuroactive compounds from microelectrode array recordings of cortical neural networks. 2020 Annual Meeting of the Society of Toxicology, Anaheim, California, March 15 - 19, 2020. https://doi.org/10.23645/epacomptox.19104635
Impact/Purpose:
Poster presented to the Society of Toxicology annual meeting March 2020. Screening chemicals for acute effects on neural network function will be valuable to OPPTS under the Lautenberg Chemical Safety Act
Description:
Human exposure to environmental chemicals can result in acute neurotoxicity (NT), negatively impacting brain activity for short and/or long periods of time. In vitro microelectrode array (MEAs) recordings of neural network function following chemical exposure can be used to screen chemicals for NT hazard. These recordings capture temporal (from minutes to days) and spatial aspects of action potential activity as described by a set of network parameters. To date, more than 1055 compounds have been screened in this assay, including testing 384 compounds at different concentrations following 40 mins exposure. To determine if a compound is neuroactive, global network parameters are extracted from the entire 40 mins of neural recordings resulting in loss of temporal information (TI). The TI could improve identification of compound fingerprints and/or could provide information on the mechanisms of action that mediate a neural network response after acute exposure of a compound. Here, data from 384 previously tested compounds were used to explore the properties of the TI to screen for acute neuroactive compounds using the response from a single high concentration (nominally 40 µM). A window analysis technique was applied. From previous baseline and post-dosed recordings on day in vitro (DIV) 12, a total of 19 network parameters were extracted for each window of time of 1 minute with overlaps of 30 seconds, resulting in one time series (trajectory) per parameter. Extracted parameter trajectories were normalized per well and per window of time, and a moving median filter was applied to reduce outliers. A k-means clustering technique was used to find 10 clusters of trajectories for each parameter and used to assign cluster IDs to the trajectories of each compound. Then for each compound, a vector with a total of 19 cluster IDs (one per parameter) was assigned and used to classify the compounds into neuroactive or negative compounds using a Support Vector Machine (SVM) classifier. The entire classification model was trained with 73 compounds (42 neuroactives, 31 negatives) and yielded a classification accuracy of 93.2%. When using the model to classify the 384 compounds, 257 were identified as neuroactives and 127 as negatives. By comparison, when TI was excluded from the SVM classifier, classification accuracy of the same 73 neuroactive/negative compounds decreased to 86.3%. The higher classification accuracy of the SVM model that uses data with TI demonstrates the value of TI for identifying acute neuroactive compounds more effectively when performing single-point screening. This abstract does not reflect policy of the US EPA.
URLs/Downloads:
DOI: Semi-supervised Classification Algorithm using k-means Trajectory Clustering and Support Vector Machine to identify Neuroactive compounds from microelectrode array recordings of cortical neural networks
SOT 2020 POSTER SYLMARIE GRIP_V4.PDF (PDF, NA pp, 701.925 KB, about PDF)