Adverse Childhood Experiences: The Hidden Force Behind So Many Behaviours

How AI, Trauma Screening, and the Positive Systems Approach May Change the Way We Help Children

by Dr. Bob Carey

One of the greatest risks in working with children is believing that we already know why they are behaving the way they are. A child who cannot sit still is labelled as having ADHD. A student who refuses to complete schoolwork is described as oppositional or unmotivated. Another who withdraws from classmates is viewed as shy, anxious, or socially awkward. These descriptions may contain elements of truth, but they do not always tell the whole story. Sometimes the behaviour we see is not the problem at all. It is the symptom.

Over the past several decades, researchers have developed an increasingly sophisticated understanding of Adverse Childhood Experiences (ACEs)—events such as abuse, neglect, domestic violence, parental mental illness, substance misuse, chronic conflict, or the loss of a caregiver. These experiences can profoundly influence a child's developing brain, emotional regulation, relationships, learning, and physical health. Importantly, the effects of childhood trauma often extend well into adulthood. What is becoming increasingly clear is that trauma rarely announces itself openly. Instead, it often disguises itself as something else.

A child living with unresolved trauma may appear inattentive, impulsive, anxious, depressed, aggressive, withdrawn, perfectionistic, or emotionally explosive. Others simply seem disengaged, tired, or uninterested in school. In many cases, these behaviours resemble conditions such as ADHD, anxiety disorders, mood disorders, learning disabilities, or autism spectrum disorder. Sometimes those diagnoses are accurate. Sometimes trauma is the primary driving force. Often, both are present. This complexity reminds us why understanding behaviour requires curiosity before conclusions.

Recent advances in artificial intelligence (AI) and digital health technologies are beginning to offer new ways of recognizing children who may be struggling long before their distress becomes obvious. Rather than replacing psychologists or other mental health professionals, these tools are being designed to help identify patterns that might otherwise go unnoticed.

Researchers are developing systems capable of integrating information from questionnaires, electronic health records, school attendance, sleep patterns, wearable devices, and even subtle changes in language or behaviour over time. Together, these sources of information may help identify children who are at increased risk for developing post-traumatic stress disorder (PTSD), depression, anxiety, or even suicidal thinking. Unlike traditional assessments that provide a snapshot at one point in time, digital health tools can monitor changes continuously. A gradual decline in sleep quality, increasing social withdrawal, more frequent absences from school, and subtle shifts in mood may all occur weeks or months before a child receives a referral for psychological services. AI systems are becoming increasingly capable of recognizing these patterns and alerting clinicians that further assessment may be warranted.

Perhaps one of the most encouraging findings from this emerging research is that many young people disclose emotional distress more readily through digital platforms than they do during face-to-face conversations. For adolescents in particular, digital screening may lower some of the barriers created by stigma, embarrassment, or fear of disappointing adults. Yet, despite the excitement surrounding these technologies, we must also proceed thoughtfully. Artificial intelligence cannot understand a child's life story. It cannot appreciate the meaning of a parent's divorce, the experience of chronic bullying, the death of a grandparent, or the daily stress of living in poverty. It cannot witness the warmth of a caring teacher, the comfort of a supportive grandparent, or the resilience that develops through positive relationships. AI can identify statistical patterns, but it cannot fully understand human experience.

This is where the role of the psychologist remains irreplaceable. Technology may help us recognize children who need help earlier than ever before, but understanding why those children are struggling—and how best to help them—still requires compassionate clinical assessment, thoughtful listening, and genuine human connection. This is also where the Positive Systems Approach offers an important perspective.

One of the central principles of the Positive Systems Approach is that behaviour never exists in isolation. Every behaviour occurs within a larger system of relationships, environments, experiences, and biological influences. Rather than asking, "How do we stop this behaviour?" we begin by asking a different question:  "What is this behaviour trying to communicate?" When viewed through this lens, behaviour becomes valuable information rather than something to simply eliminate.  A child who refuses school may not simply be oppositional. A child who lashes out at classmates may not simply lack self-control. A teenager who appears depressed may not simply need greater motivation. Their behaviour may represent an attempt to cope with overwhelming stress, unresolved trauma, sensory overload, anxiety, or feelings of emotional insecurity.  The Positive Systems Approach encourages us to examine the entire ecology surrounding the child. We consider family relationships, school expectations, peer interactions, community supports, physical health, developmental history, and previous life experiences. Trauma is not viewed as a separate issue but as one of many interacting influences that shape behaviour over time.

Importantly, the approach also reminds us to look beyond risk and recognize strengths.  Children who have experienced adversity often demonstrate remarkable resilience. They survive difficult circumstances by developing coping strategies that once served an important purpose. Some become hypervigilant because vigilance helped keep them safe. Others become emotionally guarded because vulnerability once carried significant risk. These adaptations are understandable responses to difficult environments, even if they later interfere with healthy development.  The goal is not to judge these behaviours but to understand them.  This systems perspective also complements the emerging role of AI remarkably well.

Artificial intelligence excels at identifying patterns across large amounts of information. The Positive Systems Approach excels at understanding the meaning behind those patterns. Together, they have the potential to improve both early identification and effective intervention. Imagine a future in which digital screening tools identify a child whose sleep has deteriorated, whose school attendance has begun to decline, and whose mood questionnaires reveal increasing distress. Rather than simply generating another referral, the psychologist uses that information as the starting point for a comprehensive assessment that explores trauma history, family functioning, developmental experiences, school climate, protective factors, and the child's own understanding of what is happening.

Technology helps us know where to look.  The Positive Systems Approach helps us understand what we are seeing.

Perhaps this is one of the most exciting opportunities emerging in children's mental health today. We are moving away from reacting only after crises occur and toward recognizing vulnerability much earlier. Earlier recognition creates opportunities for earlier support, stronger relationships, and more effective prevention.

As clinicians, educators, and parents, we should welcome innovations that help us notice children who might otherwise remain invisible. At the same time, we must remember that no algorithm can replace empathy, curiosity, or the healing power of human relationships.  We need to remember that children are more than their symptoms…..they are more than their diagnoses and they are certainly more than their behaviours. When we combine thoughtful clinical judgment, trauma-informed care, emerging technologies, and a systems-based understanding of behaviour, we move closer to seeing the whole child—not simply the behaviour that first caught our attention.

Key Points at a Glance:

Current State of the Evidence

A major 2024 narrative review by Brianna M. White and colleagues examined digital health technologies and AI for children exposed to ACEs. They reviewed studies involving digital screening, machine learning, natural language processing (NLP), mobile health applications, and AI-assisted decision support.

Their overall conclusion was cautiously optimistic:

• Digital health technologies can improve early detection of mental health consequences following ACEs.

• AI can identify patterns that clinicians may overlook.

• Digital monitoring can identify changes before symptoms become severe.

• Ethical safeguards, privacy protections, and clinician oversight remain essential.

White BM, Prasad R, Ammar N, Yaun JA, Shaban-Nejad A. Digital Health Innovations for Screening and Mitigating Mental Health Impacts of Adverse Childhood Experiences: Narrative Review. JMIR Pediatr Parent. 2024 Oct 16;7:e58403. doi: 10.2196/58403. PMID: 39412745; PMCID: PMC11498064.

What Can AI Detect?

PTSD (Strong Emerging Evidence)

Research is strongest for PTSD.

AI systems have demonstrated the ability to detect PTSD risk using:

• electronic health records

• trauma questionnaires

• speech analysis

• written language

• smartphone behaviour

• physiological monitoring (heart rate, sleep)

Machine learning models analyze hundreds of variables simultaneously and often outperform traditional risk scoring methods.

Examples include:

• language reflecting hypervigilance

• sleep disruption

• withdrawal from peers

• changes in phone activity

• emotional language

• school attendance patterns

Rather than diagnosing PTSD, AI estimates the probability that a child is developing trauma-related symptoms. Several pilot studies report prediction accuracies between approximately 75% and 90%, although performance varies depending on the population studied and the quality of the training data. These findings require further validation in real-world pediatric settings.

White BM, Prasad R, Ammar N, Yaun JA, Shaban-Nejad A. Digital Health Innovations for Screening and Mitigating Mental Health Impacts of Adverse Childhood Experiences: Narrative Review. JMIR Pediatr Parent. 2024 Oct 16;7:e58403. doi: 10.2196/58403. PMID: 39412745; PMCID: PMC11498064.

Depression (Strong Evidence)

Depression is probably the best-studied application.

AI models successfully integrate:

• PHQ questionnaires

• mood ratings

• sleep data

• social interaction patterns

• physical activity

• electronic medical records

• school attendance

• language analysis

These systems can identify gradual worsening months before many children seek treatment.

Digital monitoring is particularly valuable because depressive symptoms often fluctuate.

Repeated assessment provides far more useful information than a single clinic visit.

Several systematic reviews now conclude that digital monitoring improves recognition of depression, particularly in adolescents.

Fernández-Batanero JM, Fernández-Cerero J, Montenegro-Rueda M, Fernández-Cerero D. Effectiveness of Digital Mental Health Interventions for Children and Adolescents. Children. 2025; 12(3):353. https://doi.org/10.3390/children12030353

Anxiety Disorders (Strong Evidence)

Similar findings exist for anxiety.

AI detects:

• avoidance behaviours

• excessive reassurance seeking

• sleep disturbance

• elevated physiological arousal

• changes in smartphone usage

• increased isolation

Some digital platforms combine:

• wearable sensors

• ecological momentary assessment (EMA)

• brief daily symptom check-ins

• passive behavioural monitoring

This allows clinicians to identify worsening anxiety in near real time.

Studies consistently report improvements in screening accuracy compared with relying solely on periodic questionnaires.

Fernández-Batanero JM, Fernández-Cerero J, Montenegro-Rueda M, Fernández-Cerero D. Effectiveness of Digital Mental Health Interventions for Children and Adolescents. Children. 2025; 12(3):353. https://doi.org/10.3390/children12030353

Types of Digital Health Technologies Being Studied

Current research includes:

Technology	Current Use
Mobile apps	Mood tracking, symptom screening
AI chatbots	Initial mental health screening, psychoeducation
Machine learning	Risk prediction from multiple data sources
Natural Language Processing	Analysis of speech, interviews and written responses
Digital phenotyping	Passive monitoring through smartphones
Wearables	Sleep, activity and physiological monitoring
Electronic Health Record AI	Identification of high-risk children in healthcare systems
School-based digital screening	Universal screening in educational settings

Why AI May Be Especially Helpful After ACEs

Children with ACEs often present with overlapping symptoms.

For example - Trauma may resemble:

• ADHD

• anxiety

• depression

• autism

• oppositional behaviour

• learning problems

Traditional assessments often examine these conditions separately.

Machine learning can instead integrate information across multiple domains simultaneously, including:

• trauma history

• behaviour

• family risk

• medical history

• educational performance

• emotional symptoms

• social functioning

This systems-level approach aligns well with contemporary trauma-informed care.

Major Advantages

Current research consistently identifies several benefits.

Earlier identification

Children are often recognized months earlier than through routine referrals.

Continuous monitoring

Instead of one annual assessment, symptoms can be monitored daily or weekly.

Increased access

Digital tools can reach rural communities and underserved populations.

Reduced stigma

Many adolescents disclose sensitive symptoms more readily through digital interfaces than during face-to-face interviews.

Precision risk prediction

AI can combine hundreds of risk factors simultaneously.

Important Limitations

The literature is equally clear about current limitations.

AI does not diagnose.

It estimates probability.

Clinical diagnosis remains the responsibility of trained professionals.

Algorithmic bias

Many AI models were trained on relatively homogeneous populations.

Performance may be poorer among:

• Indigenous youth

• racialized communities

• culturally diverse populations

• children with developmental disabilities

Privacy

ACE information is among the most sensitive health information collected.

Researchers consistently emphasize:

• encryption

• informed consent

• parental involvement (when appropriate)

• transparent governance

• secure data storage

Over-reliance

Children should never receive treatment decisions based solely on AI.

Clinician judgement remains essential.

Where the Field Is Heading

Research published in 2025 suggests the next generation of systems will integrate multiple data streams into a single clinical dashboard. These platforms may combine:

• ACE screening questionnaires

• wearable sensor data

• smartphone-based digital phenotyping

• electronic health records

• school attendance and performance

• caregiver reports

• clinician observations

The goal is to provide clinicians with a continuously updated estimate of risk rather than relying on isolated assessments. At the same time, reviews emphasize the need for rigorous validation, youth-centered design, privacy safeguards, and clear crisis-response protocols before widespread implementation.