Detecting Probable Alzheimer's Disease From Speech Using Linguistical Analysis

Detecting Mild Cognitive Impairment and Probable Alzheimer's Disease From Speech Using Linguistic Deficits With Amyloid PET Imaging as a Baseline.

The object of this study is to investigate the use of linguistic deficits from speech samples for the early detection of Mild Cognitive Impairment and probable Alzheimer's disease. It will also evaluate whether the result of the Amyloid PET scan would confirm the effectiveness of a less expensive and less intrusive diagnostic technique through speech

This study will investigate the use of linguistic deficits through a computational diagnostic model for the prediction of Mild Cognitive Impairment (MCI) and Probable Alzheimer's disease (PrAD) from speech samples (verbal utterances). The premise is that MCI and PrAD are characterized by the deterioration of nerve cells that control cognitive speech and language processes, which consequently affects how patients compose verbal utterances. As the Amyloid Positron Emission Tomography (PET) scan is making advances in giving confirmatory evidence of the existence of plaques in the brain, there is the opportunity to investigate whether the result of the scan would confirm the effectiveness of a less expensive and less intrusive diagnostic technique. In this study, participants of a previous IRB approved study on Imaging Dementia-Evidence for Amyloid Scanning (IDEAS) who consent to participate in the new study will be requested to perform a basic descriptive task. The main task to be done by the participants is to verbally describe the scenes on a projected image, while their verbal utterances will be recorded via an audio recorder. The descriptive task is expected to take between 10 to 15 minutes on average. Participants will be required to describe the scenes on the projected Cookie-Theft image (http://www.amyspeechlanguagetherapy.com/thecookie- theft-picture.html). The image is part of the Provide a brief but thorough description of the study. What is the study about? This does not need to contain a summary of background literature. It should just summarize the study itself. Boston Diagnostic Aphasia Examination (BDAE-3) and it is adapted in our study for the purpose of capturing the complexity in the linguistic expression similar to the DementiaBank project (https://talkbank.org/media/DementiaBank/Pitt/cookie/). Note that we are NOT conducting any psychological examination using the BDAE-3, we have only referenced the source of the Cookie-Theft image, which is publicly available for research and has been used for decades. An example of the task is demonstrated in this video: https://www.youtube.com/watch?v=1NB8wBBguMg. The transcripts of the recorded interview will be analyzed by our proposed computational model to detect the presence of Mild Cognitive Impairment or probable Alzheimer's disease. Finally, the IDEAS study PET scan data would be used in conjunction with the transcripts of the recorded interview to compare the results of the linguistic analysis of the transcript with those of the PET scan in making Mild Cognitive Impairment or probable Alzheimer's Disease diagnoses.

Inclusion Criteria:

• Participants who have undergone the Mini Mental State Examination (MMSE).
• Participants who have undergone Amyloid PET scan with a diagnosis of having the Amyloid plaques or not.
• Participants with moderate or fluent English speaking ability.

Exclusion Criteria:

• Participants who cannot speak as the data collection process needs to obtain speech samples.
• Participants who are severely demented or at a very late stage of Alzheimer's disease.
• Participants with stroke.
• Participants who report a history of other neurodegenerative disorders than MCI and AD.
• Participants with chronic vision impairment or who cannot see with the aid of corrective glasses.
• Participants who cannot speak English.
• A participant who have been diagnosed with HIV/AIDS as this may have a cognitive effect on its own.