The progression of Alzheimer’s Disease (AD), which is the most common form of dementia, can be slowed down if detected early. Several screening tests have been developed for this purpose, MMSE (Mini-Mental State Examination) being one of the most widely used. However, such tests need to be administrated by a specialized therapist and typically consist of frequent meetings between patients and thera- pists, which difficulties their applicability. This thesis aims to develop an automatic web-based tool that enables patients to perform screening tests at home without the assistance of therapists and, simultaneously, help therapists in diagnosing and monitoring the disease. The tool will integrate the L2F in-house speech and language technologies (SLT). The present document will focus on the description of the most popular neu- ropsychological tests that involve speech components, and the existing automated tests that resort to SLT. It will also include a brief overview of an existing architecture which provides a convenient framework for the tool. The document concludes with the presentation of the planned calendar for development and the evaluation methods for the final results.
In European countries, the ratio of elderly population is increasing, particularly those countries in Western Europe, the society is aging with the shrinking portion of the younger generation. While most of the elderly people spend most of their time living at home, the importance of Research and Development of new technologies adapted to the elderly is becoming vital. As a purpose of increasing the older people's autonomy and independence, speech appears to be one of the most natural and efficient modality. The purpose of the present document is to draw the state-of-the-art in human-machine interaction via speech for the elderly, with a major focus on Automatic Speech Recognition (ASR). Ageing is responsible for a significant decrease in ASR performance because of speech production alterations due to age. Adapting the acoustic and language statistical models still is a hot research topic.
Aphasia is an acquired communication disorder that affects speech and language functionalities at varying degrees. The recovery of lost communication functionalities is possible through frequent and in- tense therapy sessions. In this context, VITHEA, an on-line platform that exploits Speech and Language Technologies (SLT) has been developed to facilitate the recovery process of Portuguese aphasic patients. This report presents the state of the art of the existing system and proposes some possible lines of research that will be beneficial to improve it.