Advancing the detection of early dementia with digital speech markers
Pedro Pinheiro-Chagas, Professor
UC San Francisco
Closed. This professor is continuing with Fall 2024 apprentices on this project; no new apprentices needed for Spring 2025.
Alzheimer’s disease and related dementias are neurodegenerative diseases with numbers rapidly increasing and currently no cure. To accurately identify the earliest signs of clinical dementia, there is a critical need for sensitive, low-cost, and high-access cognitive markers in
the preclinical phase. Novel cognitive markers can complement biomarker information to understand heterogeneity in clinical presentations, aid clinical dementia diagnosis and prognosis, and provide sensitive outcome measures for clinical trial research. Technological innovations now allow us to leverage connected speech analysis as a digital measure in dementia research, which aligns seamlessly with the imperative to develop innovative and objective cognitive markers. Connected speech analysis using Natural Language Processing (NLP) becomes pivotal in capturing the nuances of language changes, which are inadequately addressed by conventional language and cognitive tests. This project focuses on developing digital speech markers from connected speech and language samples in Alzheimer’s disease and frontotemporal dementia. Analyses will utilize advanced statistics (e.g., Structural Equation
Modeling), graph theory, and machine-learning techniques to investigate the relationships and classification ability of these digital speech markers with neuroimaging biomarkers, other cognitive tasks, and clinical diagnosis.
Role: Successful candidates will play a role in multiple aspects of the project. Responsibilities will include combining data from internal servers, audio clipping, quality checking transcriptions, data cleaning and preprocessing, coding and code checking, aiding analyses, and preparing results for dissemination. An important part of your role will be ensuring the secure and ethical handling of sensitive patient data, adhering to the HIPAA standards. In addition to the Principal Investigator (Vonk), the candidate will work closely with two data analysts in her group. This position offers a unique opportunity to apply the candidate’s skills and abilities in NLP, SEM, graph theory, and machine-learning within a healthcare context.
Qualifications: Ideal candidates should have a background in data science, computer science, biostatistics, neuroscience, epidemiology, computational linguistics or neurolinguistics, or a related field (open to other fields if the fit is right), with a good understanding of Python or R (at least one out of two), and some knowledge of NLP, machine learning, SEM, or graph theory (at least one out of four). Most importantly, we value candidates who show enthusiasm and are up for a hands-on approach to gain experience in all the steps of the research process.
Day-to-day supervisor for this project: Professor Jet Vonk, UCSF
Hours: to be negotiated
Biological & Health Sciences Digital Humanities and Data Science