Hi friend,

It was a busy week so let’s get straight into it.

⏺️ Ampa raised a new round. Ampa aims to make TMS treatment more accessible to clinics and patients through their custom device, Ampa One. Today they announced a new funding round that “brings their total funding to $25M”. Based on their previous fundraising ($8.5M) we can assume this round was approx. $17M. Morningside Ventures, Trimera Capital, and Tim Ferriss were all part of the investment. In their press release, Ampa said that demand for their product has been doubling every quarter. (link)

⏺️ Sonomind raised a €20M Series A for focused ultrasound depression treatment. The French company is developing a focused ultrasound system for treatment-resistant depression. The round was led by Critical Path Ventures and Bpifrance. Neuromodulation is having a moment it seems. (link)

⏺️ Ksana Health was awarded an ARPA-H contract (worth up $17.9M) to build a behavioural health AI model. The Oregon-based company will work with Providence Health & Services, MedStar Health, and the University of Washington to build a model aimed at earlier detection of mental health and substance use disorders. The contract includes $5.1M in base funding with additional options for scaled phases. (link)

⏺️ UK government launched a call for evidence for their 10-year mental health reform plan. The consultation seeks input from clinicians, service providers, and people with lived experience, and is open until 10 July. The strategy aims to shift the system toward earlier intervention and prevention and to take a cross-governmental approach to promoting positive mental health. (link)

⏺️ Headspace launched a revamped Apple Watch experience. The update brings guided microinterventions of breath-work, mindfulness and sleep content onto the Apple Watch in easier to use formats. The new app will also use Apple Health data — specifically HRV — to identify moments where users might be more open to engaging in activities. On a personal level, I’ve become very interested in both breath-work and HRV so I’m excited to see where this goes — now I just need an Apple Watch... (link)

⏺️ MindBench released a pilot of the MindBench Community Benchmark. This benchmark scores frontier models on two separate questions: how well a model’s ratings align with the consensus of clinicians and people with lived experience, and whether they correctly identify responses that the community defined as harmful. The initial results, leaderboard and methodology are available on the MindBench website. (link)

⏺️ MindBench also released a Model Cards feature, making it easy to track what each AI labs’ model card says about their approach to mental health safety testing. (link)

⏺️ Sword Health's mental health AI product Dawn surpassed 100,000 people on its waitlist. Sword have been pushing Dawn hard in recent months with significant out of home marketing campaigns across New York (yes, I’ve seen the Subway and Time Square ads). I’m waiting to see how Dawn actually performs and what they do with all these users who’ve signed up… (link)

⏺️ In the UK, the health insurer Vitality launched a new children's mental health offering. Mental health support will now be available to children on policyholders' plans and will include services delivered by Headspace, Wysa and Mindler. (link)

⏺️ Prospira Global published a framework for integrating mental health into global development programmes. The report makes the case for embedding mental health as a structural priority within funder portfolios rather than treating it as a standalone programme area. The central argument is that mental health is a cross-cutting determinant of whether sectoral investments succeed, and that while recognition of this is growing, financing structures and measurement frameworks have not kept pace. Worth a read. (link)

⏺️ Wellcome launched a $1M prize for mental health science breakthroughs. The Wellcome Prize for Mental Health Science, run in partnership with Nature, will award $1M to the overall winner and $250K each to three finalists. The prize aims to accelerate effective innovations but also to shine a light on the breakthroughs in this space to get policymakers, funders and the general public excited about what the future of mental healthcare could look like. It is open to research teams and small and medium-sized organisations worldwide, and will recognise advances in treatments for anxiety, depression, and psychosis that are scientifically rigorous and show potential to scale. This will be the world’s largest prize for mental health science. Pretty cool. (link)

If you find these updates valuable and would like to connect with other leaders in mental health innovation, consider joining the Hemingway Community.

As a member of the community, you’ll join over 450 other mental health leaders and get access to exclusive content, events and resources, including our upcoming Session with Derrick Hull PhD, R&D Lead at Slingshot AI.

⏺️ Nature Mental Health review argues depression research has a measurement problem. Despite decades of study, depression research has delivered limited progress in improving predictive models and treatment outcome. The review argues the problem is how depression is defined and measured — with too much focus on aggregate scores, insufficient attention to how symptoms change over time, and almost no input from patients on what outcomes matter to them. The authors propose focusing on specific symptom domains like anhedonia and sleep, tracking trajectories over time, and building patient-defined outcomes into research design. (link)

⏺️ Population-based RCT finds digital CBT self-help reduced depression in college students. Pairing population-wide mental health screening with a digital CBT guided self-help program was found to reduce the prevalence of anxiety, depression, and eating disorders significantly more than a referral-to-care comparison. Benefits persisted at two years and service uptake more than doubled, including among minority groups. (link)

⏺️ Scoping review of AI language models in mental health finds limitations in real-world validation. Researchers looked at 41 studies published between 2023 and 2025 to understand how AI language models like ChatGPT are being used in mental health care. They found that these tools are mostly being tested for diagnostic use cases, and while many studies showed promising results, fewer than one in three actually tested the AI with real patients and against real clinicians. (link)

⏺️ Network meta-analysis of prenatal exercise finds moderate aerobic training best for depression, yoga best for anxiety. Researchers analysed 36 RCTs involving over 4,500 pregnant women to compare how different types and intensities of exercise affect mental health during pregnancy. They found that moderate-intensity aerobic exercise was most effective at reducing prenatal depression, while yoga was best for lowering anxiety and regulating the stress hormone cortisol. Exercising three times a week for more than 12 weeks produced the largest benefits overall. (link)

⏺️ CBT outperformed ACT for OCD patients with high anxiety and emotion regulation difficulties. A study of 176 OCD patients found that those with high anxiety sensitivity and difficulty regulating emotions responded better to CBT than to ACT. (link)

⏺️ 37% of Australian high school students report having experienced suicidal ideation. That’s a truly terrifying number. The Orygen study found suicidal ideation to be far more common than most population estimates suggest. (link)

⏺️ Orygen researchers also published a major review on the role of social media in preventing suicide and self-harm. The review found that social media platforms are increasingly used for suicide and self-harm prevention at scale — through both formal interventions and user-led practices like peer support — but while evidence supports their safety and acceptability, research remains limited to short-term outcomes and the field lacks the longitudinal evidence needed to guide policy. (link)

⏺️ Bipolar disorder linked to accelerated biological ageing. ECG data from 278 bipolar disorder patients showed a 2.5-year gap between physiological and chronological age. The finding points to cardiovascular risk mechanisms in bipolar disorder. (link)

⏺️ Lower choline levels found in the brains of people with anxiety disorders. Researchers found reduced choline levels in the prefrontal cortex of anxiety disorder patients, pointing to a possible biochemical basis for anxiety. Practical therapeutic applications remain unclear at this stage. (link)

⏺️ No reliable predictors found for ketamine response in treatment-resistant depression. A systematic review of 122 studies found no consistent clinical variables that predicted whether a patient would respond to ketamine or esketamine. (link)

⏺️ Therapist-rated factors may help predict who responds to psychedelic-assisted therapy. A Nature Mental Health study found that certain therapist ratings made during sessions were associated with treatment outcomes in psychedelic-assisted therapy. The study found that therapeutic alliance, social support, personality openness, and the capacity to surrender were the strongest predictors of positive outcomes, while prior non-psychedelic substance use was seen as the most unfavourable predictor. Meaningful differences were found between between psilocybin and ketamine therapists, and between regulated and unregulated practice settings. (link)

⏺️ RCT finds single psilocybin dose reduces major depression rapidly, but effects fade by 12 months. In a placebo-controlled trial of 35 adults with recurrent major depression, a single 25mg dose of psilocybin produced significant symptom reduction within two days and maintained superiority over placebo through six weeks. The advantage was no longer statistically significant at 12 months however. (link)

That’s all for this week.

Keep fighting the good fight!

Steve Duke

Founder of Hemingway

P.S. Find this valuable and want to support our work? Consider becoming a Hemingway Pro member.