Dr Ross Harper, Co-founder and CEO of Limbic talks about his journey to starting Limbic, a company that provides AI software for mental healthcare.


The evolution of Limbic


Limbic is a product of both my academic background and my Co-founder Sebastiaan’s background in software development.


After studying natural sciences at Cambridge University, I became interested in applying mathematical modelling to studies of the brain. I therefore embarked on a Postgraduate Masters in Mathematical Modelling at University College London and then went on to do a PhD in Computational Neuroscience. I began to believe that the solutions to some of the biggest challenges in mental illness would come from statistical and computational approaches. 


This journey, coupled with the fact that some of the biggest scientific innovations were happening outside the traditional lab and university settings, caused me to join the Entrepreneur First accelerator programme, rather than continue my path in academia. I joined the programme with an idea for an AI-based approach to mental healthcare where I met my Co-founder Sebastiaan, a programming whiz. 


Solving the supply and demand mismatch


Globally, one in four people will be affected by mental or neurological disorders at some point in their lives. Around 450 million people currently suffer from such conditions, placing mental health disorders among the leading causes of ill-health and disability worldwide.


However, there are growing concerns about the number of staff available to deliver mental health support services, largely due to the low number of available therapists. Fundamentally, we have a supply and demand mismatch which has only been exacerbated by the pandemic. Limbic is designed to bridge this gap. 


Our platform uses conversational AI to provide end-to-end support to clinicians and patients at different points in their treatment pathway, from referral through to post-recovery, all while collecting data that is forwarded on to clinicians to help them construct a clinical profile. Ultimately, this can mean increased clinical efficiencies, reduced patient wait times and improved patient outcomes.. 


For example, our self-referral tool, Limbic Access, allows patients to refer into mental health support services, rather than rely on clinicians to do this manually. Not only does this save time – 1214 clinical hours to date in fact – but it also cuts the patient wait time. In the NHS-based talk therapy services we work with, patients wait 22 days (on average) for an assessment. Limbic has the potential to cut this down to zero. In our pilots, we were able to save a total of 3867 weeks of patient wait time. 


Once the patient has completed their assessment, they are usually placed on a second wait list. In a typical setting, they wait an additional 45 days (on average) to commence treatment. However, our mobile app, Limbic Self-Care, provides instant CBT support to patients. The data collected during this time is forwarded on to the clinician, which helps to speed up the first treatment session..


Once the patient is assigned a therapist and face-to-face treatment begins, the platform transitions into Limbic Care, a remote monitoring app that is coached by the therapist to provide personalised tips to the patient in-between sessions – altering the conversational AI in real-time. Post-treatment, the platform transitions to Limbic Prevent, which offers continued personalised coping strategies, to mitigate the risk of relapse, an issue that  currently affects more than half of mental health patients.​​​​​​​


Left: Limbic Access, Right: Limbic Self-Care


What we’ve learnt 


Our platform is constantly evolving and we’ve faced a number of challenges. While AI has been identified as a way to bridge the gap between supply and demand, psychological therapy is a very human-centric discipline. Machine learning has been adopted in many other areas of healthcare but mental health is different – arguably, it requires the biggest human touch. A challenge for our team has been striking the right balance between integrating digital solutions and preserving the human aspect of psychological therapy. 


This is why our team believes that full automation in this space is naive. We focus less on replacing therapists, and more on augmenting them. Our software seeks to solve specific pain points in the clinical pathway. By doing this, we are able to keep healthcare fundamentally human and personalised, while also freeing up clinical hours wherever it doesn’t make sense to use valuable human resources. We often think of it as giving the clinician superpowers. 


We are only just beginning to see the ramifications of the pandemic on mental health services. Digital solutions must focus on empowering clinicians and helping them make data-driven decisions, while extending their teachings to the patient in-between sessions. The result will be increased service capacity and enhanced care at a time when mental health services need it most.


 The World Health Report 2001: ‘Mental Health: New Understanding, New Hope’, [accessed 8 July 2021], https://www.who.int/whr/2001/en/whr01_en.pdf 


MixPanel data as of 7 July 2021 


NHS Digital: Waiting times for referrals entering treatment in the year 2019-20: https://app.powerbi.com/view?r=eyJrIjoiMGUwOGFiYWItODFhYS00MmMzLWFkMjQtOTg4NzU4MDk4ZTI3IiwidCI6IjUwZjYwNzFmLWJiZmUtNDAxYS04ODAzLTY3Mzc0OGU2MjllMiIsImMiOjh9


MixPanel data as of 7 July 2021 


NHS Digital: Waiting times for referrals entering treatment in the year 2019-20: https://app.powerbi.com/view?r=eyJrIjoiMGUwOGFiYWItODFhYS00MmMzLWFkMjQtOTg4NzU4MDk4ZTI3IiwidCI6IjUwZjYwNzFmLWJiZmUtNDAxYS04ODAzLTY3Mzc0OGU2MjllMiIsImMiOjh9


 * Ali, et al, “How durable is the effect of low intensity CBT for depression and anxiety? Remission and relapse in a longitudinal cohort study;” Behaviour Research and Therapy, V 94, July 2017, [Accessed 19th April 2021] https://www.sciencedirect.com/science/article/abs/pii/S0005796717300840?via%3Dihub