Using our unique Natural Processing technology the Persyst Seizure Detector simulates the way that human reviewers scan for seizure patterns. It takes input from the Rhythmicity Spectrograms as well as a number of other factors in order to spot the kind of changes that represent seizure activity. Through its unique ability to recognize electrical and chewing artifacts it avoids the types of false positives that have made seizure detectors less effective in the past. The result is an overall accuracy rate of nearly 90% with only one to two false positives per twenty-four hours when tested against our extensive internal database of EMU records. Of course results with individual patients can vary which is why we have uniquely integrated our seizure detection trends with our overall trending solution to provide the most complete picture available.
The Rhythmicity Spectrogram is a key input to the Persyst Seizure Detection algorithm. It measures the amount of rhythmicity which is present at each frequency in the record.