Commentary

AI could play role in treatment of mental health, neurological disorders

Image of Guillermo Cecchi
Guillermo Cecchi

by Guillermo Cecchi, PhD, MSc

Artificial Intelligence and machine learning are becoming more prevalent in health care. The science of training computers is providing clinicians with additional insights into the mental health of patients, which they may be able to soon use to determine appropriate treatment plans.

The arsenal of data that AI gives us may be the key to managing two of America’s fastest growing epidemics: mental health and neurological disorders.

The prevalence of mental health and neurological disorders continues to rise in the U.S. Nearly one in every five people in the U.S. has a mental illness, and one in every six people in the U.S. has a neurological disorder. However, a shortage of professionals dedicated to treating these diseases remains an ongoing and growing issue. By 2025, it's estimated that the demand for psychiatrists may outstrip the supply of psychiatrists by up to 15,600. This is a problem, especially given that for these conditions, the earlier a diagnosis is made, the higher the likelihood is that it can be successfully treated and managed by health care professionals.

As part of our ongoing work in the emerging field of computational psychiatry, the team at IBM Research is focused on finding a solution to the question, “How can we help the doctors and patients who are impacted by neurological diseases?” Better understanding the complexities of the human brain with science and technology is a mission we have undertaken for the past two decades, and our recent advances in brain and disease progression modeling with AI are giving us clues.

Since one of the earliest indicators of cognitive decline can be changes in speech, we have found the answer may lie with AI. By using computational biology, analytics and machine learning, we can build tools which can help clinicians to quickly and simply analyze language and better predict the onset of these diseases. Being able to diagnose these conditions rapidly and simply may allow for earlier intervention, better allocation of resources and better treatment planning.

This advancement is also important because, typically, language analysis is a manual process that can take months to conduct. Then, even once the data have been collected, there is often not a standard for processing the information and turning it into an informative diagnosis. So, while changes in speech are known to provide solid clues into mental health, it’s previously been difficult for clinicians to capture and analyze these data in an efficient and meaningful way. This is where AI comes in.

By taking short samples of speech input, systems can use text-to-speech, advanced analytics, machine learning, natural language processing technologies and computational biology to quickly provide an overview of the patient’s mental health.

This technology has been tested with patients who have either mental health or neurological disorders. In a 2015 study conducted with Columbia University, we were able to predict, with 100% accuracy, which individuals among a population of at-risk adolescents would develop their first episode of psychosis within 2 years. Three years later, in collaboration with academic institutions, researchers were able to better track, predict and monitor disease in Parkinson’s patients using snippets of speech.

In early 2017, we predicted that in 5 years, what we say and write will be used as indicators of our mental health and physical well-being — patterns in our speech and writing analyzed by cognitive systems will provide results of early-stage mental disease that prompt us to seek treatment.

Not even 2 years later, this is moving more toward becoming a reality.

In addition to the research with Columbia and Pfizer, recent progress has been made in helping doctors measure mental and neurological health with speech and provide clues that point to possible dementia, schizophrenia and psychotic episodes.

AI algorithms have been used to help inform clinicians about a user's neurological state based on the structural complexity of their sentences.

AI and machine learning techniques have helped clinicians identify critical language patterns and utilize machine learning to quickly analyze verbal speech and alert clinicians of the possibility of a mental episode.

Together, we are transforming how we can approach mental health. As our work and this field of research as a whole progresses, we will likely soon be able to develop technology and succinct techniques to provide answers to this health epidemic we are currently facing.

For more information:

Guillermo Cecchi, PhD, MSc, is research staff member, Computational Neuroscience, at IBM Research.

References:

Bedi G, et al. npj Schizophrenia. 2015;1:doi:10.1038/npjschz.2015.30. eCollection 2015.

Cecchi G, Corcoran CM. Schizophr Bull. 2018;44(Suppl 1):S76.

Corcoran CM, et al. World Psychiatry. 2018;doi:https://doi.org/10.1002/wps.

DiLuca M, et al. Bull World Health Organ. 2018;96:298–298A.

Eyigoz E, Cecchi G, Tejwani R. Proceedings of the 10th International Conference on Agents and Artificial Intelligence (ICAART 2018). 2018;2:683-692.

National Council for Behavioral Health. The Psychiatric Shortage: Causes and Solutions March 28, 2017. Available at: https://www.thenationalcouncil.org/wp-content/uploads/2017/03/Psychiatric-Shortage_National-Council-.pdf. Accessed March 10, 2019.

Substance Abuse and Mental Health Services Administration. Key Substance Use and Mental Health Indicators in the United States: Results from the 2015 National Survey on Drug Use and Health. Available at: https://www.samhsa.gov/data/sites/default/files/NSDUH-FFR1-2015/NSDUH-FFR1-2015/NSDUH-FFR1-2015.pdf. Accessed March 11, 2019.

UN News. Nearly 1 in 6 of world’s population suffer from neurological disorders – UN report. Available at: https://news.un.org/en/story/2007/02/210312-nearly-1-6-worlds-population-suffer-neurological-disorders-un-report. Accessed March 11, 2019.

Disclosure: Cecchi is an employee at IBM Research.

Image of Guillermo Cecchi
Guillermo Cecchi

by Guillermo Cecchi, PhD, MSc

Artificial Intelligence and machine learning are becoming more prevalent in health care. The science of training computers is providing clinicians with additional insights into the mental health of patients, which they may be able to soon use to determine appropriate treatment plans.

The arsenal of data that AI gives us may be the key to managing two of America’s fastest growing epidemics: mental health and neurological disorders.

The prevalence of mental health and neurological disorders continues to rise in the U.S. Nearly one in every five people in the U.S. has a mental illness, and one in every six people in the U.S. has a neurological disorder. However, a shortage of professionals dedicated to treating these diseases remains an ongoing and growing issue. By 2025, it's estimated that the demand for psychiatrists may outstrip the supply of psychiatrists by up to 15,600. This is a problem, especially given that for these conditions, the earlier a diagnosis is made, the higher the likelihood is that it can be successfully treated and managed by health care professionals.

As part of our ongoing work in the emerging field of computational psychiatry, the team at IBM Research is focused on finding a solution to the question, “How can we help the doctors and patients who are impacted by neurological diseases?” Better understanding the complexities of the human brain with science and technology is a mission we have undertaken for the past two decades, and our recent advances in brain and disease progression modeling with AI are giving us clues.

Since one of the earliest indicators of cognitive decline can be changes in speech, we have found the answer may lie with AI. By using computational biology, analytics and machine learning, we can build tools which can help clinicians to quickly and simply analyze language and better predict the onset of these diseases. Being able to diagnose these conditions rapidly and simply may allow for earlier intervention, better allocation of resources and better treatment planning.

This advancement is also important because, typically, language analysis is a manual process that can take months to conduct. Then, even once the data have been collected, there is often not a standard for processing the information and turning it into an informative diagnosis. So, while changes in speech are known to provide solid clues into mental health, it’s previously been difficult for clinicians to capture and analyze these data in an efficient and meaningful way. This is where AI comes in.

PAGE BREAK

By taking short samples of speech input, systems can use text-to-speech, advanced analytics, machine learning, natural language processing technologies and computational biology to quickly provide an overview of the patient’s mental health.

This technology has been tested with patients who have either mental health or neurological disorders. In a 2015 study conducted with Columbia University, we were able to predict, with 100% accuracy, which individuals among a population of at-risk adolescents would develop their first episode of psychosis within 2 years. Three years later, in collaboration with academic institutions, researchers were able to better track, predict and monitor disease in Parkinson’s patients using snippets of speech.

In early 2017, we predicted that in 5 years, what we say and write will be used as indicators of our mental health and physical well-being — patterns in our speech and writing analyzed by cognitive systems will provide results of early-stage mental disease that prompt us to seek treatment.

Not even 2 years later, this is moving more toward becoming a reality.

In addition to the research with Columbia and Pfizer, recent progress has been made in helping doctors measure mental and neurological health with speech and provide clues that point to possible dementia, schizophrenia and psychotic episodes.

AI algorithms have been used to help inform clinicians about a user's neurological state based on the structural complexity of their sentences.

AI and machine learning techniques have helped clinicians identify critical language patterns and utilize machine learning to quickly analyze verbal speech and alert clinicians of the possibility of a mental episode.

Together, we are transforming how we can approach mental health. As our work and this field of research as a whole progresses, we will likely soon be able to develop technology and succinct techniques to provide answers to this health epidemic we are currently facing.

For more information:

Guillermo Cecchi, PhD, MSc, is research staff member, Computational Neuroscience, at IBM Research.

References:

Bedi G, et al. npj Schizophrenia. 2015;1:doi:10.1038/npjschz.2015.30. eCollection 2015.

Cecchi G, Corcoran CM. Schizophr Bull. 2018;44(Suppl 1):S76.

Corcoran CM, et al. World Psychiatry. 2018;doi:https://doi.org/10.1002/wps.

DiLuca M, et al. Bull World Health Organ. 2018;96:298–298A.

Eyigoz E, Cecchi G, Tejwani R. Proceedings of the 10th International Conference on Agents and Artificial Intelligence (ICAART 2018). 2018;2:683-692.

National Council for Behavioral Health. The Psychiatric Shortage: Causes and Solutions March 28, 2017. Available at: https://www.thenationalcouncil.org/wp-content/uploads/2017/03/Psychiatric-Shortage_National-Council-.pdf. Accessed March 10, 2019.

Substance Abuse and Mental Health Services Administration. Key Substance Use and Mental Health Indicators in the United States: Results from the 2015 National Survey on Drug Use and Health. Available at: https://www.samhsa.gov/data/sites/default/files/NSDUH-FFR1-2015/NSDUH-FFR1-2015/NSDUH-FFR1-2015.pdf. Accessed March 11, 2019.

UN News. Nearly 1 in 6 of world’s population suffer from neurological disorders – UN report. Available at: https://news.un.org/en/story/2007/02/210312-nearly-1-6-worlds-population-suffer-neurological-disorders-un-report. Accessed March 11, 2019.

Disclosure: Cecchi is an employee at IBM Research.