Artificial intelligence isn’t just showing up in behavioral health—it’s beginning to rewire how we detect, diagnose, and deliver care. Three shifts are worth watching: predictive diagnostics that surface risk earlier (sometimes before symptoms are obvious), objective tech-enabled assessments that reduce subjectivity and support ongoing progress monitoring, and digital therapeutics and virtual platforms that expand reach—especially in communities that have historically been underserved.

Below are two companies illustrating how this is unfolding, plus what it could mean for clinicians, families, and payors.

Predictive Diagnostics: LinusBio and the “Exposome” in a Strand of Hair

What they do: North Brunswick, NJ-based LinusBio (a Mount Sinai spinout) is building diagnostic tools designed to estimate a child’s likelihood of autism by analyzing molecular “fingerprints” left by environmental exposures (the exposome) captured along a single strand of hair. One centimeter corresponds to roughly a month of exposure history; using laser-based analysis, the platform reconstructs a time-series map of biomarkers to power AI models.

The company’s flagship product, StrandDx-ASD, received FDA Breakthrough Device designation in December 2021. In February 2025, LinusBio launched ClearStrand-ASD, a commercially available “rule-out” test offered through their CLIA-certified laboratory. ClearStrand-ASD provides a 92.5% negative predictive value—meaning a negative result indicates a child is unlikely to be on the autism spectrum—and is now available in 49 U.S. states plus Japan. (Note: ClearStrand-ASD has not been FDA cleared or approved; it is a lab-developed test.)

Why it matters: LinusBio’s thesis is simple: if we can quantify risk earlier, we can route families to evaluation and intervention sooner—where timing often shapes outcomes. Autism has largely been identified through behavioral observation. ClearStrand-ASD aims to augment, not replace, that process with biology-informed risk insights.

Momentum: Following the 2021 Breakthrough designation, LinusBio raised $16 million in January 2023 to expand R&D and clinical validation. In April 2025, the company announced a partnership with Autism Speaks to advance early diagnosis and intervention, including offering ClearStrand-ASD at reduced cost to families in need.

Reality check: ClearStrand-ASD is an aid, not a stand-alone diagnosis. Clinical validation and real-world utility data are essential before wide adoption. Ethical guardrails matter—especially around communication of risk, consent, and follow-up pathways to evaluation and services.

Objective, Earlier Identification & Ongoing Monitoring: EarliPoint Health’s Eye-Tracking

What they do: EarliPoint Health (formerly EarliTec Diagnostics, rebranded October 2025) uses patented eye-tracking and visual attention biomarkers to provide objective, quantifiable insights related to early autism indicators and to support ongoing assessment during treatment. The FDA-cleared platform is designed to be fast, standardized, and friendly for young children ages 16 to 30 months.

Why it matters: Objective tasks can help flag risk in toddlers, supporting referral for comprehensive evaluation sooner. For clinicians, repeatable, standardized signals can help monitor progress and tailor care plans.

Recent developments: In April 2024, EarliTec raised $21.5 million in Series B financing led by Nexus NeuroTech Ventures and Venture Investors Health Fund. In August 2025, Jamie Pagliaro—a longtime operator in autism services and ed-tech who previously served as COO at Rethink—became President & CEO, signaling a push into scaling access and commercialization. The October 2025 rebrand to EarliPoint Health coincided with the launch of the EarliPoint Network, a new model designed to connect families with trusted provider partners who serve as evaluation locations.

Context: EarliPoint Health joins other FDA-cleared digital tools aimed at earlier autism detection (e.g., AI-enabled apps, eye-tracking approaches). The shared theme: reduce friction, reduce subjectivity, and shorten the path from concern to care.

Digital Therapeutics & Virtual Care: Expanding the “Last Mile”

AI is also showing up in care delivery—triaging families to the right level of care, supporting parent-mediated interventions, and assisting with documentation and outcomes tracking. For rural and underserved areas, this can mean fewer waitlists and more tailored supports (asynchronous coaching, shorter sessions, or stepped-care models). The key is pairing tech with clinician oversight and clear outcome measures.

What This Could Mean—For Clinicians, Families, and Payors

For clinicians: If predictive and early-signal tools reduce time to evaluation, clinicians can prioritize the highest-risk cases sooner. Objective measures (like attention biomarkers) can inform data-driven treatment plans and adjust intensity and dosage. AI can also cut documentation burden and surface insights from the noisy stream of session data.

For families: When concerns arise, families need fast, reliable pathways—not just screenings but actionable next steps. Objective repeat measures can help families understand what’s changing over time.

For payors and systems: Earlier identification plus tiered, data-informed care can reduce long waitlists and optimize intensity. Standardized measures help distinguish signal from noise and align reimbursement with value.

What to Watch (and What to Ask)

Validation and equity: Do models perform similarly across demographics, languages, and settings? Are false positives and negatives acceptable and clearly communicated?

Clinical workflow fit: Can tools be run in minutes, with clean handoffs to evaluation and services? Do they reduce burden rather than add it?

Data governance: Who owns the data? How are results stored, shared, and explained? What happens after a positive screen?

Real-world outcomes: Do these tools actually shrink time to diagnosis, improve access, and enhance outcomes—not just produce impressive AUCs (area under the curve, a statistical measure of diagnostic accuracy)?

Reimbursement: Are there clear coding pathways and payer policies to support sustainable use?

Bottom Line

AI won’t replace the clinician—but clinicians who leverage AI may outperform those who don’t. LinusBio is pushing on the earliest signal—turning hair into a high-resolution exposure timeline that may guide faster evaluation. EarliPoint Health is standardizing what we measure—offering objective, repeatable indicators to support early identification and track progress.

The opportunity is compelling: earlier insights, more personalized care, and broader reach. The responsibility is clear: validate rigorously, implement ethically, and keep humans in the loop.