Multifold features determine linear equation for automatic spike detectionapplying neural nin interictal ECoG

Objective: A 3-layer detection procedure was designed including preselectio n applying TEMPLAS software, feature extraction and artificial neural netwo rks to determine a fast, precise and highly selective spike algorithm. Methods: Ten intraoperative ECoG recordings of patients with temporal lobe epilepsy were computer-assisted and evaluated by 3 experts upon preselected events. For each event, 19 features were extracted, normalized and fed int o a two-layer and 3-layer feedforward, back-propagate network. The weights of the 5 best individual two-layer networks of patients were averaged separ ately to derive a mean network, where weights were pruned, rounded off and the configuration approximated by a linear equation. Results: In addition. when investigating latency histograms, a method for m ulti-channel artefact detection and elimination of too close intra-channel events could be found. Out of several training trails only the mean network and the linear equation were able to generalize. In comparison with the re sults of 19 publications, the developed solution and the estimated overall detection rates (spikes: 81%; non-spikes: 99.3%) were found to be of high q uality. The processing time is shea, and therefore, the method can be used to initiate other measurements. Conclusion: The developed solution is a fast, precise and highly selective spike detection method. (C) 1999 Elsevier Science Ireland Ltd. All rights r eserved.