Acoustic signal processing applications for the support of dementia sufferers

This research project studies the use of acoustic signal processing techniques for the support and care of dementia sufferers. More specifically, it aims to investigate how wearable microphone array technologies can be used to assist patients affected by dementia.

Microphone array processing techniques exploit multiple microphone signals to try and extract a desired signal, such as speech, from a noisy environment containing various types of interferences. Using wearable microphone arrays as an assistive technology for dementia sufferers leads to two main challenges. First, many of the currently existing acoustic array processing techniques employ static arrays, in which the constituent microphones do not move. However, using wearable arrays means that the constituent microphones' positions in space are not always known and they may move relative to one another. For example, microphones sewn in a jumper or t-shirt move in space and relative to each other as the human wearer interacts with its environment or as the clothe stretches or shrinks over time. Thus, novel signal processing algorithms must be developed to account for uncertainty in the array configuration.
Then, aiming to develop acoustic processing algorithms to assist a specific user group, in this case dementia sufferers, means that their hearing impairments must be identified, quantified, and appropriately described. Thus, this project must identify which aspects of current state-of-the-art acoustic processing techniques are inappropriate for dementia patients, which are the most likely to have a large impact on their quality of life, and how the performance of algorithms can be evaluated in a way that faithfully translates the perception of dementia sufferers. Novel evaluation measures must therefore be developed that can properly represent the performance of array processing techniques in assisting dementia sufferers.

Solving hearing impairments experienced by dementia sufferers is important to ensure they receive a proper diagnosis and suitable care. Indeed, dementia can cause hearing problems that are related to the accurate comprehension of sounds, such that the dementia syndrome often appears to be worse than in reality. For example, certain dementia sufferers may lose the ability to extract spatial information from sounds. This can cause disorientation in noisy or busy environments and lead medical professionals to assume the patient is in a very frail state, when in fact their cognition may still be largely unaffected. Since dementia is often irreversible, the best treatment for patients is often highly individualised, and requires the true extent of the disease to be known. Thus, any factor such as hearing impairment masking the true cognitive state of patients must be treated in order to ensure the patients receives the best possible care.

This project aligns with the Digital Signal Processing EPSRC research area and shows potential applications to the Clinical Technologies research area