The Study
EverythingALS is working with clinical research stakeholders from IBM Research, Massachusetts Institute of Technology, and Harvard University in a longitudinal study focused on early disease detection based on patterns noted in audiovisual sessions.
The study includes over 800 participants diagnosed with Amyotrophic Lateral Sclerosis (ALS), caregivers, those who are suspected to develop ALS based on family history, and genetic carriers for ALS.
EverythingALS, a patient-focused non-profit, part of the Peter Cohen Foundation, wants to enroll more participants to accelerate the study in meeting its goal of finding a pain to successfully diagnosing ALS. According to John Hopkins Medicine, ALS affects as many as 30,000 people in the United States, with 5,000 new cases diagnosed each year.
While the study’s focus in on ALS, health care partners are hoping to use the AI technology to find better ways to diagnose and treat Parkinson’s, Alzheimer’s, and other neurological conditions.
Participants in the speech research study complete short audiovisual sessions of the course of the year. The protocol is computer-based and includes participant tasks such as reading sentences out loud for an audiovisual recording.
The goal of the protocol is to use the study partner’s AI technology to spot and track early signs of ALS and assess progression of ALS through speech and facial gestures.
Mindy Uhrlaub, advocate, author, and genetic carrier of ALS, participated in the study for a year. She submitted audiovisual samples weekly over a year-long period. Uhrlaub described the research program as “studying your mastery of language but also the mastery of your mouth and ability to breathe.”
The study is accessible through a free online application created by Modality.ai which allows participants like Uhrlaub easy access to the study without having to travel which is often difficult for ALS patients. Uhrlaub explained that the protocol of the study often includes repeating back phrases and the practice of doing it once a week became part of her routine.
All of the data from the research conducted is open data that researchers, pharmaceutical companies, and potential partners have access to and are invited to collaborate on.
Reversing the “backwards” diagnostic process
Founder of EverythingALS, Indu Navar, started the non-profit after losing her husband, Peter Cohen, to ALS. Navar explained that her husband didn’t receive his ALS diagnoses until roughly two and half years after he began exhibiting symptoms of the disease. At the point of diagnoses, Navar’s husband had already experienced roughly 60% of the damage the disease would incur.
Navar felt that the method of diagnoses for ALS was “backwards” as the current protocol for ALS diagnostics is through a process of elimination. According to Johns Hopkins Medicine, there is not one test or procedure to ultimately establish the diagnosis of the disease. Therefore, the disease can be left untreated and progressing before the diagnoses takes place.
There is also no known cure for ALS and on average, according to John Hopkins Medicine, individuals with the disease live three to five years after diagnoses.
With her background in technology, Navar was focused on working on a study that connected raw materials to finished products. In this case, the study connects the raw data of the participants speech patterns and connects it to the product of digital biomarkers created by the audiovisual analysis created by AI.
Leveraging AI for future tools
IBM Research explained that the company’s artificial intelligence (AI) system analyzes the voice and video to determine digital biomarkers.
Through the digital biomarkers that can be established through this study, researchers may be able to determine changes on a granular level that can be applied to diagnoses and future clinical trials for treatments designed to slow ALS.
Norel told us that because patients often visit doctors at a lower frequency then once a week there is often a very clear decline in an ALS patient’s abilities and increased progression of their symptoms. However, the data collected from this study may be able to create a tool to help clinicians see “more subtle changes” in a patient without having to go into a clinic or doctor’s office.
She believes the digital biomarkers created from this trial can be leveraged into a tool used for early detection among other potential uses.
