How Precision Behavioral A.I. Will Save Psychedelic Clinical Trials: Lessons from the FDA's Rejection of MDMA Therapy for PTSD

The FDA’s MDMA Rejection: A Wake-Up Call for Psychedelic Drug Companies.

In June 2024, the U.S. Food and Drug Administration (FDA) panel made a surprising decision to reject MDMA therapy for post-traumatic stress disorder (PTSD).

Despite promising results, the FDA cited concerns over study design, lack of objective measures, and participant bias, leading to an overwhelming vote against its approval. This decision underscores the need for new technologies, such as precision behavioral A.I., to address these challenges. If integrated into clinical trials, behavioral A.I. could have provided more objective data, potentially leading to a different outcome.

This is only the first. As the psychedelic drug market expands, many other companies are likely to face similar obstacles.

This is a big concern for me.

I believe that these medicines are especially important for trauma and the existential distress that comes with terminal diseases and end of life care. After getting diagnosed with cancer, I was allowed to participate in a small clinical trial at the Huntsman Cancer Institute in which 4 cancer patients received high doses of psilocybin and were taken through 8 hours of mind bending care guided by emotion provoking music.

For me, it was magical and healing. I felt a deep connection to the world and to my family and gained an understanding and perspective on life. I saw connections between generations that helped to heal my fears for my kids. By the end, I hoped many patients would get to have such an important experience. Indeed, people often rate these psychedelic experiences as one the most valuable of their lives. It can be a treasure, if done correctly. It can be a mess, if done poorly.

Fundamentally, any effective psychedelic treatment must be paired with the right cognitive behavioral therapy to process the experience (integration). The patient must be in the right state of mind to have a positive outcome (intention setting). This is all hard to standardize in the existing system.

Luckily, technology can help.

At Storyline (https://storylinehealth.com) , we have developed the world's leading precision behavioral A.I. technology capable of objectively measuring facial, vocal, and speech patterns to enhance the evaluation of cognitive and emotional responses in clinical settings. The use of such technologies in MDMA trials could have addressed many of the FDA’s concerns, providing more reliable and objective data to measure outcomes, as well as creating a reproducible and standardized clinical care experience.

Subjective clinical and patient reported outcomes are a known problem in clinical trials and FDA authorization (1-6). Subjective measures and bias are major factors in why clinical trial results don’t translate for patients (7). For example, rating scales are commonly used to measure different symptoms or quality of life factors, but they are problematic because we do not really know which variables most rating scales measure (8). Additionally, the crude and overly broad diagnostic criteria provided by the DSM-V has long been seen as problematic (9-10). This article explores how behavioral A.I. must become an essential new part of every clinical trial. These tools will change the trajectory of MDMA’s approval and its potential impact on the rapidly growing psychedelic drug industry.

The Psychedelic Drug Market: A Booming Industry Facing Challenges

The global psychedelic drug market is projected to grow significantly in the coming years, with estimates suggesting it could reach $10.75 billion by 2027. Dozens of companies are now involved in the development of psychedelic-based therapies for mental health conditions such as PTSD, depression, anxiety, and substance use disorders. Companies like Compass Pathways, MindMed, atai Life Sciences, and GH Research plc are actively advancing clinical trials for compounds like #psilocybin, #LSD, and #DMT.

Each of these companies is at different stages of regulatory approval, but all face the same scrutiny from regulatory bodies, particularly in terms of study design, participant bias, and the subjective nature of measuring psychiatric outcomes.

The rejection of #MDMA therapy by the #FDA raises red flags for other companies in this space. As more psychedelic drugs move through the clinical pipeline, developers must overcome the same issues that derailed MDMA’s approval: the difficulty of blinding studies, potential placebo effects, and a lack of standardized measures for assessing therapeutic efficacy. Precision Behavioral A.I. could play a pivotal role in helping these companies generate more reliable, objective data, addressing the concerns that currently impede their path to market.

The Role of Precision Behavioral A.I. in Addressing Bias and Subjectivity

One of the core challenges in psychedelic trials is the inability to blind participants and therapists effectively. In the MDMA trials, both groups could often tell whether they had received the drug or a placebo, leading to possible placebo effects and biased interpretations of outcomes. This is especially problematic when the primary and secondary outcomes involve subjective clinical and patient reports.

I have written previously about the challenges of open label trials and how the FDA struggles with subjective patient reported outcome (PRO) data in these trial designs due to measurable biases compared to objective data types. Subjective data types block progress in many fields and clinical trial designs. Clearly, a treatment that improves the patient’s reported quality of life is important, yet it is just a fact that not all therapeutic interventions fit the randomized, double blind, placebo controlled model.

Precision Behavioral A.I. helps to solve this issue by providing objective measures of patient symptomatic, behavioral, emotional and cognitive phenotypes. By analyzing tens of thousands of micro-features that capture complex behavior, vocal tones, and speech patterns from smartphone-captured video, Storyline A.I. systems objectively measure shifts in behavior, mood, and stress levels due to treatment.

Trials should be designed with a primary endpoint, secondary endpoint that can involve clinical PRO questionnaires, and tertiary endpoints that are exploratory and use Storyline to perform deep phenotyping and collect objective measurements that test the hypothesis and support the primary and secondary endpoints. Collectively, this body of data goes to the FDA to support the review for authorization.

For instance, PTSD patients undergoing MDMA therapy often experience emotional breakthroughs, which are typically measured through self-reporting or therapist observations. Behavioral A.I. could supplement these subjective measures by objectively tracking emotional shifts in real-time during therapy sessions. This data would provide regulators with more concrete evidence of the drug’s impact, reducing the reliance on potentially biased self-reports and improving confidence in the results.

Addressing Safety Concerns with Continuous Monitoring

Safety is another critical issue raised by the FDA, particularly in psychedelic drug trials. Many participants in the MDMA trials had previously used illicit MDMA, which could have biased the results. Furthermore, the long-term effects of MDMA use, especially when combined with other treatments during the follow-up period, were not fully understood. Behavioral A.I. could enhance the safety profile of such trials by providing continuous, objective monitoring of participants' emotional and psychological states. Storyline’s deep phenotyping symptom assessment takes only 10-20 seconds per day by a patient and collects tens of thousands of data points.

Through real-time analysis of vocal stress indicators, facial expressions of discomfort, and cognitive speech patterns, A.I. could detect early signs of adverse effects, enabling earlier intervention. This continuous data stream would allow researchers to create more personalized safety protocols and offer regulators a clearer picture of both the benefits and risks associated with the therapy.

Standardizing Psychotherapy in Clinical Trials

A key issue raised by the FDA panel was the variability in psychotherapy administered alongside MDMA. The FDA noted that it was difficult to separate the effects of the drug from the therapist’s influence, as therapists in the trial had significant discretion in their approach. Behavioral A.I. could help standardize the measurement of psychotherapy effectiveness by objectively analyzing the interactions between therapists and patients.

Our technology at Storyline Health is developing A.I. agents capable of delivering reproducible, standardized Cognitive Behavioral Therapy (#CBT) via a smartphone. These adaptive algorithms integrate our deep phenotyping measures of symptoms and behavior with custom Large Language Models (LLMs) and decision models that return appropriate CBT questions and exercises for the patient to perform on their smartphone in a video interview. This algorithmic approach enables a reproducible strategy to provide CBT reproducibly, adaptive CBT for each patient, and implements decision tools that help optimize the outcome for each patient by delivering the right CBT intervention. This approach standardizes the therapy across clinical trial sites and in the care setting following approval.

Storyline Behavioral AI CBT smartphone exercise - “Create a personal story around your struggles, focus on how you've grown, what you've learned, and how you can use this knowledge in your life.”

In addition to standardizing the therapy experience across trial sites and care facilities, by measuring subtle changes in a patient's response to different therapeutic interventions—such as changes in tone, phrasing, or body language—A.I. provides valuable insights into the consistency and effectiveness of therapy across different trial settings. This would help regulators like the FDA distinguish the drug's effects from the influence of the therapeutic environment, offering a more reliable assessment of the drug’s true efficacy.

Finally, the Behavioral AI assessments, CBT interventions, and decision algorithms are powerful companion products that can be optimized in trials and coupled with the drug in the market to increase revenue. This has a somewhat similar regulatory framework as a companion biomarker or digital health product.

The Future of Psychedelic Clinical Trials

The FDA’s rejection of MDMA therapy for PTSD highlights the broader challenges faced by the psychedelic drug industry. With over 20 psychedelic compounds currently in development across more than 80 clinical trials, companies are racing to bring these therapies to market. However, unless they can provide more objective, data-driven evidence of safety and efficacy, many of these drugs may face the same regulatory hurdles as MDMA.

By integrating behavioral A.I. into trial design, developers can enhance their data collection and analysis, addressing many of the concerns that have traditionally hampered the approval of psychedelic therapies. Objective measurements of emotional and cognitive changes could replace or supplement subjective self-reports, giving regulators a more robust foundation on which to base their decisions.

Conclusion

As the psychedelic drug market continues to expand, the need for objective, standardized measures in clinical trials has never been more pressing. Behavioral A.I. has the potential to revolutionize how these therapies are evaluated, offering more precise and reliable data to regulators. Had such technology been used in the MDMA trials, the outcome might have been different, potentially paving the way for faster, safer approval of psychedelic therapies.

We are at the forefront of this technological revolution, developing A.I.-driven tools that will transform the evaluation of mental health therapies. As psychedelic treatments become more mainstream, behavioral A.I. will be essential in ensuring their successful, evidence-based integration into the medical landscape.

REFERENCES:

1. Roydhouse, J. K., Fiero, M. H. & Kluetz, P. G. Investigating Potential Bias in Patient-Reported Outcomes in Open-label Cancer Trials. JAMA Oncol. 5, 457–458 (2019).

2. Wood, L. et al. Empirical evidence of bias in treatment effect estimates in controlled trials with different interventions and outcomes: meta-epidemiological study. BMJ 336, 601 (2008).

3. Savović, J. et al. Influence of Reported Study Design Characteristics on Intervention Effect Estimates From Randomized, Controlled Trials. Ann. Intern. Med. 157, 429 (2012).

4. Gnanasakthy, A., Barrett, A., Evans, E., D’Alessio, D. & Romano, C. (DeMuro). A Review of Patient-Reported Outcomes Labeling for Oncology Drugs Approved by the FDA and the EMA (2012-2016). Value Heal. 22, 203–209 (2019).

5. Jahedi, S. & Méndez, F. On the advantages and disadvantages of subjective measures. J. Econ. Behav. Organ. 98, 97–114 (2014).

6. Moustgaard, H., Bello, S., Miller, F. G. & Hróbjartsson, A. Subjective and objective outcomes in randomized clinical trials: definitions differed in methods publications and were often absent from trial reports. J. Clin. Epidemiology 67, 1327–1334 (2014).

7. Heneghan, C., Goldacre, B. & Mahtani, K. R. Why clinical trial outcomes fail to translate into benefits for patients. Trials 18, 122 (2017).

8. Hobart, J. C., Cano, S. J., Zajicek, J. P. & Thompson, A. J. Rating scales as outcome measures for clinical trials in neurology: problems, solutions, and recommendations. Lancet Neurol. 6, 1094–1105 (2007).

9. Insel, T. R. & Cuthbert, B. N. Brain disorders? Precisely. Science 348, 499–500 (2015).

10. Cuthbert, B. N. & Insel, T. R. Toward the future of psychiatric diagnosis: the seven pillars of RDoC. BMC Medicine (2013).

#CBT #PrecisionBehavior #BehavioralAI #ClinicalTrials #PrecisionMedicine #PsychedelicResearch #FDAapproval #FDA #PrecisionMedicine #DrugResearch #DigitalHealth #HealthcareAI

Christopher Gregg PhD is a tenured Professor in the departments of Neurobiology and Human Genetics at the University of Utah School of Medicine , and the Chief Science Officer and Co-founder of Storyline Health, Primordial AI, and Uncharted Cancer Patient Masterclass.

A leader in genomics, evolution, metabolism, data science technologies, and computational analyses of natural behavior. Recipient of the Eppendorf & Science Prize in Neurobiology, the NYSCF Robertson-Neuroscience Investigator award. Work selected as a top breakthrough of the year by the NIMH in 2010 and again in 2018 by STAT. Committed to developing precision medicine solutions that radically improve patient care.

Disclosure: Chris Gregg has a cofounder and has financial interests in Storyline Health Inc., and Primordial AI Inc.