Hi friend,

2026 continues to be a big year for Neuromodulation. This week Wave received new FDA clearance and both Brainsway and Nexalin announced strong new outcomes data.

In policy news, the UK goverment announced today that they plan to ban social media (and intimate AI companions) for under 16s.

But that wasn’t all… Talkspace released a new AI Chatbot, InStride raised $30M and new interesting research showed that smart plug data can be highly predictive of mental health risk in cancer survivors.

There’s lots more happening too, so let’s get into it.

We’ve got lots of events coming up for Hemingway Pro members…

NYC Chapter: The Long Table — Wednesday, June 17th: A new Hemingway format hosted by our NYC Chapter lead Jane Rho. We will be hosting a few small dinners running in parallel across the city, each with a different mental health theme. Sign up here.

Sydney Chapter: Hemingway Mornings — Thursday, June 25th: Sydney Chapter Lead Richard Eve is hosting a small structured breakfast for 8 at Paramount Coffee Project. Sign up here.

Hemingway DC: Happy Hour — Thursday, June 25th: Come along for a Hemingway happy hour at The Golden Age. Sign up here.

Chicago: Potluck Dinner — Friday, June 26th: Come and join the Chicago crew for a casual potluck dinner. The wonderful Ben Anadappa is hosting! Sign up here.

Hemingway Session — Tuesday June 30th: Beyond the Crisis Narrative: Building Mental Health Solutions That Actually Scale with Ben Locke. Sign up here.

SF Chapter: The Local — Tuesday June 30th: If you're based in SF, join Lukas Wolf as he kicks off the first “SF Local”. More details to come soon so keep an eye on the SF Chapter Slack channel.

London Chapter Launch — Wednesday 1st July: We are officially launching Hemingway London! Come along for some pints, good people, and conversation at a proper London pub. Sign up here.

Hemingway Session — Thursday July 9th: For Youth, By Youth: How to Build Mental Health Products That Young People Actually Use with Jas Schembri-Stothart, co-founder of Luna. Sign up here.

⏺️ InStride Health raised $30M to expand pediatric mental health services. Echo Health Ventures and FMZ Ventures led the series C round. InStride Health provides insurance-based virtual treatment for anxiety and OCD in patients aged 7 to 24, and has raised $86M to date. The company operated in 17 states, up from 8 in 2024, and said it had treated 5,000 patients. (link)

⏺️ Nonprofit QCF/I will acquire Advantage Behavioral Health. QCF/I is a nonprofit that owns and operates addiction treatment and mental health facilities and they are set to acquire New Jersey-based Advantage Behavioral Health, Advantage operates 48 locations in New Jersey and sites in seven other states. Advantage had received an investment from private equity firm Clearview Capital in April 2025. (link)

⏺️ Active Minds received $5M from Google.org to expand their youth mental health programmes. The funding will support digital wellbeing and mental health programming for 100,000 youth and young adults across 154 high schools, with a focus on digital safety skills, peer support networks, and youth advocacy. (link)

⏺️ SAMHSA announced $40M in funding across eight grant programs. The Substance Abuse and Mental Health Services Administration announced funding of $9.2M for behavioural health and community safety partnerships, $9M for Tribal Behavioral Health substance use prevention, $8M each for school and primary care screening programmes and for the National Center for Child Traumatic Stress. Smaller awards covered adult suicide prevention, a centre of excellence for eating disorder training, peer-run mental health organisations, and addiction treatment education for graduate health students. (link)

⏺️ Talkspace launched a direct to consumer AI companion, Tee. Tee is a direct-to-consumer AI mental health companion priced at $19.99 a month. The product will sit alongside, rather than within, Talkspace's existing therapy and psychiatry services, and can refer users to human clinicians when needed. You can my analysis and commentary on the announcement here. (link)

⏺️ Wave Neuroscience received FDA clearance for their PTSD neuromodulation device. Wave's MeRT (or Magnetic EEG-guided Resonance Therapy) uses EEG-based analysis of patient brains to deliver personalised treatment through neuromodulation devices. The clearance followed a Breakthrough Device designation granted in 2024. (link)

⏺️ Sunflower began offering virtual access to naltrexone in Pennsylvania. Addiction startup Sunflower began prescribing naltrexone virtually for alcohol use disorder in Pennsylvania. (link)

⏺️ Black Dog Institute and Tacklit partner to deliver better blended care for youth depression. The two Australian organisations are partnering to integrate Tacklit's digital platform into Black Dog's clinical pathways in an attempt to advance blended care models for youth depression. While Black Dog have their own digital mental health program, Clearly Me, they will use Tacklit's infrastructure to facilitate the delivery of blended care and measure engagement and impact. (link)

⏺️ Allia Health launched a payer reimbursement transparency tool. The free microsite shows rates across a collection of payers, providers and reimbursement codes based on the Transparency in Coverage dataset. The data shows significant variation in rates across the market. (link)

⏺️ The UK government announced plans to ban social media for under-16s. Under-16s will be banned from social media platforms including Snapchat, TikTok, YouTube, Instagram, Facebook, and X, following the model implemented by the Australian Government last year. The Government said it would restrict livestreaming and stranger contact for under-16s and "Romantic companion" chatbots and all AI offering "intimate functionalities" will have to enforce a minimum age of 18. Legislation is expected before Parliament by the end of the year with the ban expected to come into effect in Spring 2027. Reaction amongst academics and young people has been mixed. While many agree that young people are struggling and that social media poses challenges, there is much less agreement that banning social media is an effective solution to these problems. Organisations like Wellcome are commissioning significant research on this topic in an attempt to build the evidence base. (link)

⏺️ In the UK, the Health Innovation Network of South London published a blueprint for how digital mental health technologies can support children and young people with anxiety and depression. The blueprint maps where digital tools could best provide support at each stage of a patient's journey, from referral to discharge. It's a thorough document that communicates the real challenges young people and their parents face when seeking care as well as several opportunities for where DMHTs can provide solutions. (link)

⏺️ CBT AI chatbots show small to modest effects for depression. A meta-analysis of 15 randomised controlled trials and 1,737 participants found that CBT-based NLP-enabled AI Conversational Agents (or CBT-B-NLP-E-AI-CAs for short...lol) produced small to moderate effects on depression. Effects on generalised anxiety, stress, and positive affect were not significant once the analysis adjusted for publication bias. Sub group analysis showed that multi-modal agents may outperform single-modality ones for depression, and younger users showed greater gains. (link)

⏺️ Study found NHS Talking Therapies mismatched patients. Researchers analysed data from 89,467 NHS Talking Therapies patients from 2014-2017 across three low-intensity treatments: individual guided self-help, group guided self-help, and computerised CBT. They also built a causal forest model which used each patient's baseline data to predict which treatment would have worked best for them individually. They then compared the real outcomes with the expected outcomes when the causal forest model assigned the treatment. In the real, observational data most patients were assigned to guided self-help (88%), but the model recommended that treatment for only 33.5% of patients. Instead, the model largely favouring computerised CBT, assigning it to 64% of patients. Among patients where the model was most confident, those who received its recommended treatment had an 11.3 percentage point higher recovery rate based on the model’s predicted outcomes. The data is quite old but the research raises some important questions around how treatments are assigned to patients. (link)

⏺️ Digital Clinic model found to cut high-cost healthcare visits after ED discharge. A retrospective evaluation of 79 patients discharged from Beth Israel Lahey Health-Needham's ED found that enrolment in the Digital Clinic, a hybrid programme combining brief sessions, a smartphone app, and a digital navigator, significantly reduced healthcare visits, particularly hospital admissions, ED care, and urgent care. Findings are preliminary. (link)

⏺️ Community mental health follow-up cut hospitalisations by two-thirds. A retrospective study of 652 patients with schizophrenia spectrum disorders in Türkiye found that long-term follow-up at a community mental health centre was associated with a 66% reduction in psychiatric hospitalisations. Annual admission rates fell from 0.179 to 0.061 per patient after enrolment, and longer follow-up duration was linked to further reductions. (link)

⏺️ Smart plugs found capable of predicting mental health risk in cancer survivors. Researchers conducted a 12-week study of 41 older cancer survivors using smart plugs, activity trackers, and smart scales to detect patterns linked to mental health risk. Television-viewing patterns alone classified risk with an accuracy of 0.71, while a model combining all three sensor types reached 0.85. (link)

⏺️ BrainsWay announce new research showing SWIFT Deep TMS remission to be held at 12 months. SWIFT is a TMS protocol requiring significantly less visits than standard TMS but questions around durability have always been a challenge. This new announcement from Brainsway states that their research found that 83% of patients who achieved remission with their SWIFT accelerated Deep TMS protocol remained in remission at 12 months. Furthermore, 78% achieved remission immediately after treatment and 84% at 6 months. Improvements in quality of life and disability lasted throughout the year, and fewer than 25% of patients needed medication changes or additional TMS. At the time of publshing I have been unable to source the original research paper but the findings were presented at the Annual Clinical TMS Society (CTMSS) Meeting two weeks ago. (link)

⏺️ Nexalin announce 77.8% anxiety response rate from Deep Intracranial Frequency Stimulation (DIFS) device. Nexalin makes Deep Intracranial Frequency Stimulation (DIFS) devices, which deliver low-current electrical stimulation to the brain. The company reported an 8-week, open-label trial in adults with generalized anxiety disorder and insomnia, in which Hamilton Anxiety Rating Scale scores fell 18.30 points by session 20, with 77.8% of patients responding and reductions maintained at 8-week follow-up. Depression and insomnia were also found to be improved by the treatment, with no serious adverse events. (link)

⏺️ Blood marker distinguished depression-anxiety comorbidity with high accuracy. A case-control study of 77 participants, 39 with comorbid depression and anxiety and 38 matched controls, found the metabolite S-(PGJ2)-glutathione distinguished the two groups with an AUC of 0.949. The authors noted that the sample was small and that findings should be interpreted as preliminary.(link)

⏺️ Relapse prediction models fell short of clinical use. A comment in Nature Mental Health argued that most machine learning models for predicting depression relapse failed to translate into clinical practice. The authors said this was because the models lacked external validation, real-world feasibility testing, and clear measures of clinical benefit, and proposed a framework to address these gaps. Considering the amount of research being published on using AI models to predict mental illness, this commentary is an important reminder to the field. (link)

⏺️ Case report linked chatbot use to psychotic episode. A case report in BMC Psychiatry described a man with schizophrenia who interpreted generic ChatGPT safety messages as personally targeted hidden signs. According to the case, this escalated his referential delusions and led him to stop taking antipsychotic medication. The authors recommended that clinicians routinely ask about AI chatbot use during psychiatric evaluations. (link)

⏺️ A new screening tool for borderline personality features was validated in Vietnamese adolescents. Researchers tested the BPFSC-11, a brief questionnaire measuring traits like emotional instability and identity issues, in 2,326 secondary school students. It showed strong reliability and correlated with depression and child maltreatment, supporting its use for school-based screening to identify at-risk teens. (link)

That’s all for this week.

Keep fighting the good fight!

Steve Duke

Founder of Hemingway

P.S. If you find these updates valuable, consider becoming a Hemingway Pro Member. As a Pro Member, you will get access to even deeper industry insights and will be invited to join our community of over 450 mental health leaders.

P.P.S. Excuse the lateness of this week’s email. I took some holidays and have been travelling today. Picture below of my Swedish-chocolate-fueled work on the train.