P3‐111: A return to clinical skills in the early diagnosis of Alzheimer's disease

microphone by AD patients; and iii) submitting patient recordings to signal processing analysis. Methods: Perseverators were identified in a memory clinic using a cognitive symptom questionnaire. A contact microphone was affixed to the skin above the Temporo-Mandibular Joint to record speech related vibrations during description of a pictured scene. Subjects then wore the device during the course of a normal day. Recorded signals were pre-processed to reduce background noise and remove silence periods. A variety of signal features were extracted from each recorded signal and combined using Principal Component Analysis to obtain a one-dimensional feature vector. Results: Frequently occurring patterns in the feature vector were detected using a motif discovery algorithm. Some of the recorded samples were manually analyzed and examples of the repeated sequences are presented. Conclusions: We have established the feasibility of continuous recording of bone-conducted speech, and the possibility of defining distinct patterns of perseverative speech. In the next stage of the study we aim to associate the frequency of verbal repetitions with the stage, progression and type of dementia. It is possible that the method could contribute to the assessment of disease-modifying treatments.