Speech Recognition

The software platform uses annyang! “SpeechRecognition that just works.” It performs speech recognition in multiple languages. The use of speech recognition opens the door to a wealth of cognitive testing that has previously been limited to in-person assessments. This includes verbal learning and memory tests involving immediate and delayed free recall. It also includes tests of fluency where someone lists all the words they can think of belonging to a designated category. Speech recognition also opens the doors to conducting remote cognitive assessments with people who may be unable to use a computer and rely on voice commands.

The speech recognition uses the browser’s built in engine and requires an internet connection to work. During any task using speech recognition a flashing microphone is displayed clearly indicating the user that the computer is listening. The software also keeps track of when the speech recognition is processing and only allows a user to end a trial when all speech is processed. This prevents a user from speaking then pressing the finished button before the speech processing has finished.

The speech processing returns what it hears. Comparisons to the expected responses are done by the 3C Platform. This allows inclusion of valid alternative pronunciations. For instance, the words “TEN” and “TIN” may be easily confused.

However, the software platform includes audio recordings of spoken responses allowing evaluation at a later time.