Talk to me – hearables as analysis tools: assessment of cognitive resources through incidental speech analysis
Prof. Dr.-Ing. Tanja Schultz (Computer Science)
Prof. Dr. med. Tania Zieschang (Medicine, Geriatrics, Health Sciences)
Since Hearables are by definition equipped with a microphone that is constantly within reach of the wearer, Hearable could be used to analyze the wearer’s voice. Since speaking requires complex motor and cognitive skills, abstract speech features can be used as indicators of cognitive resources. In the proposed dissertation, abstract speech features will be designed and analyzed to detect both short-term changes and long-term decline of cognitive resources. Secondary tasks and distractions will be used to find suitable features that can be automatically extracted, for example at acoustic and phonetic level (speaking rate), but also at lexical (vocabulary richness), syntactic (grammar) and semantic level (complexity). As the interception and storage of spoken communication represents a massive invasion of privacy, the intended application should not process the spoken content itself, but rather analyze abstract features that do not allow any conclusions to be drawn about what was actually said – this should remain a private matter.