{"id":18794,"date":"2024-06-25T13:05:19","date_gmt":"2024-06-25T13:05:19","guid":{"rendered":"https:\/\/interhospi.com\/?p=18794"},"modified":"2024-06-25T07:30:08","modified_gmt":"2024-06-25T07:30:08","slug":"new-ai-model-predicts-alzheimers-progression-from-speech-analysis","status":"publish","type":"post","link":"https:\/\/interhospi.com\/new-ai-model-predicts-alzheimers-progression-from-speech-analysis\/","title":{"rendered":"New AI model predicts Alzheimer\u2019s progression from speech analysis"},"content":{"rendered":"
\n

<\/p>\n<\/div><\/section><\/div>

<\/p>\n<\/div><\/section>
\n

New AI model predicts Alzheimer\u2019s progression from speech analysis<\/h1>AI<\/a><\/span>, Alzheimer\u2019s<\/a><\/span>, Boston University<\/a><\/span>, dementia<\/a><\/span>, speech<\/a><\/span>, study<\/a><\/span>, dementia<\/a>, E-News<\/a> <\/span><\/span><\/header>\n<\/div><\/section>
\n

Researchers at Boston University have developed an artificial intelligence (AI) model that can predict with 78.5% accuracy whether individuals with mild cognitive impairment will progress to Alzheimer\u2019s-associated dementia over a six-year period.<\/strong><\/p>\n

<\/p>\n

The model, which analyses speech patterns from patient interviews, could potentially revolutionise early diagnosis and intervention strategies for Alzheimer\u2019s disease.<\/p>\n

\"dementia\"<\/p>\n

The study, published in the journal Alzheimer\u2019s & Dementia<\/em> [1], utilised data from the long-running Framingham Heart Study. The research team, led by Ioannis Paschalidis, director of the BU Rafik B. Hariri Institute for Computing and Computational Science & Engineering, examined audio recordings of initial interviews with 166 participants aged 63 to 97 who had been diagnosed with mild cognitive impairment.<\/p>\n

The cohort comprised 76 individuals whose cognitive function remained stable over the subsequent six years and 90 who experienced progressive cognitive decline. Using speech recognition tools and machine learning algorithms, the researchers trained the AI model to identify correlations between speech content, demographics, diagnosis, and disease progression.<\/p>\n

Paschalidis explained: \u201cWe combine the information we extract from the audio recordings with some very basic demographics \u2013 age, gender, and so on \u2013 and we get the final score. You can think of the score as the likelihood, the probability, that someone will remain stable or transition to dementia.\u201d<\/p>\n

Implications for early intervention<\/strong><\/h3>\n

The ability to predict Alzheimer\u2019s progression with such accuracy could have significant implications for early intervention and treatment. As Paschalidis noted: \u201cIf you can predict what will happen, you have more of an opportunity and time window to intervene with drugs, and at least try to maintain the stability of the condition and prevent the transition to more severe forms of dementia.\u201d<\/p>\n

This development is particularly timely, given the recent emergence of pioneering treatments that can slow the progression of Alzheimer\u2019s disease if caught early. Currently, diagnosing Alzheimer\u2019s typically involves a comprehensive battery of assessments, including interviews, brain imaging, and blood and cerebrospinal fluid tests. However, by the time these assessments are conducted, cognitive decline may have already begun.<\/p>\n

Potential for increased accessibility<\/strong><\/h3>\n

One of the most promising aspects of this AI model is its potential to make cognitive impairment screening more accessible. By automating parts of the diagnostic process, the researchers suggest that their work could eliminate the need for expensive laboratory tests, imaging examinations, or even in-person office visits.<\/p>\n

Rhoda Au, a co-author of the study and professor of anatomy and neurobiology at BU Chobanian & Avedisian School of Medicine, emphasised the potential for AI to create \u201cequal opportunity science and healthcare\u201d. She commented: \u201cTechnology can overcome the bias of work that can only be done by those with resources, or care that has relied on specialized expertise that is not available to everyone.\u201d<\/p>\n

Limitations and future directions<\/strong><\/h3>\n

While the current model shows impressive predictive ability, the researchers acknowledge that there is room for improvement. The study relied on \u201cdirty data\u201d from casual recordings with background noise, yet still yielded significant results. This suggests that the model\u2019s performance could potentially be enhanced with higher-quality data.<\/p>\n

Paschalidis and his team are already looking ahead to future research directions. These include exploring the use of data from more natural, everyday conversations rather than formal clinician-patient interviews, and developing a smartphone app for dementia diagnosis. Additionally, they plan to incorporate other types of data from the Framingham tests, such as patient drawings and daily life patterns, to further improve the model\u2019s predictive accuracy.<\/p>\n

Conclusion<\/strong><\/h3>\n

This novel AI model represents a significant step forward in the early detection and prediction of Alzheimer\u2019s disease progression. By leveraging speech analysis and machine learning, it offers the potential for more accessible and efficient cognitive impairment screening. As Au summarised: \u201cDigital is the new blood. You can collect it, analyse it for what is known today, store it, and reanalyse it for whatever new emerges tomorrow.\u201d<\/p>\n

As research in this field continues to advance, it is hoped that such technologies will play an increasingly important role in improving early diagnosis and intervention strategies for Alzheimer\u2019s disease, ultimately leading to better outcomes for patients worldwide.<\/p>\n

Reference:<\/strong><\/h5>\n
    \n
  1. Paschalidis I., Au R., et. al. Prediction of Alzheimer\u2019s disease progression within 6 years using speech: a novel approach leveraging language models. Alzheimer\u2019s & Dementia<\/em>. 25 June 2024. https:\/\/doi.org\/10.1002\/alz.13886<\/a><\/li>\n<\/ol>\n<\/div><\/section>
    \n
    <\/span><\/span><\/div>
    \n