Lights, Camera, ADHD: Scientists Use AI to Diagnose with a Dash of GoPro Action
Thursday, December 26, 2024.
In a move that feels straight out of a sci-fi flick, researchers from the U.K. have developed a machine-learning system to diagnose ADHD by analyzing people’s actions in video clips.
Yes, diagnosing a complex neurodevelopmental disorder now involves cameras, keypads, and a little AI magic—not quite Hollywood, but definitely groundbreaking.
Published in Neuroscience Applied, this study boasts a whopping 95.5% accuracy rate, leaving traditional methods like MRI and EEG scratching their metaphorical heads.
ADHD: A Closer Look at a Chaotic Companion
ADHD (attention-deficit/hyperactivity disorder) is the frenetic DJ at the mental nightclub, spinning tracks like inattention, hyperactivity, and impulsivity.
Those with ADHD might struggle to focus, follow instructions, or organize tasks—activities that the rest of us also find difficult, but not all the time.
On top of that, there’s the signature fidgeting, restlessness, and general “can’t-sit-still-itis” that disrupts school, work, and relationships.
Most ADHD diagnoses occur when kids hit school age, where their unique quirks clash spectacularly with classroom norms.
Think of it as being asked to tango when you’ve only ever breakdanced. Unfortunately, the fallout often includes poor academic performance and strained relationships, making early diagnosis critical. Enter this cutting-edge AI system, ready to shake up the diagnostic scene.
The Science Behind the AI Show
Yichun Li and his team at Newcastle University decided it was time to bring ADHD diagnosis into the 21st century.
Their weapon of choice? Human action recognition powered by machine learning. Essentially, they set out to teach an algorithm to detect ADHD behaviors based on video footage.
Their experiment involved filming 22 participants—10 with ADHD and 12 without—using multiple GoPro cameras.
Picture a low-budget reality show where the “plot twist” is a beep reaction test instead of a dramatic argument.
Participants, aged 18 to 45, performed various activities: interviews, neurological tests, reaction challenges, and watching “exciting” videos (sadly, no details on what qualifies as “exciting”). To spice things up, distractions like fidget spinners and pens were strategically placed within reach.
The AI analyzed every fidget, glance, and touch, producing behavioral indices that screamed (or didn’t scream) ADHD. After tinkering with the data, the researchers fine-tuned their algorithm to deliver stunning results: 95.5% accuracy, outperforming traditional diagnostic tools like MRIs and EEGs, and at a fraction of the cost.
AI Diagnoses ADHD: Cheaper, Faster, Smarter
Here’s the deal. This tech isn’t just accurate—it’s cost-effective.
Traditional ADHD diagnostic tools often involve pricey equipment and lengthy processes. By contrast, this video-based system only needs a few cameras and some clever coding. According to Li’s team, the results can even assist doctors during follow-up appointments.
But, as with all trailblazing ideas, there’s a catch.
The system struggled to accurately diagnose ADHD in women.
Why?
Because women with ADHD often exhibit subtler behaviors—like “prolonged small actions”—that AI finds harder to detect. It’s a digital blind spot researchers hope to correct in future iterations.
The Fine Print
Before we hand this system a gold medal, it’s worth noting its limitations. Shorter video clips? Not great for the AI. Diagnosing ADHD in women? Still a work in progress. These hiccups remind us that while the system is impressive, it’s not perfect—yet.
ADHD Diagnosis: What’s Next?
With further refinement, this system could revolutionize how we approach ADHD diagnosis.
Imagine a world where identifying neurodevelopmental disorders is as simple as analyzing a TikTok video.
While we’re not there yet, this research is a promising leap forward, combining tech-savviness with clinical potential.
Be Well, Stay Kind, and Godspeed.
RESOURCES:
Li, Y., Nair, R., & Naqvi, S. M. (2024). ADHD Detection Based on Human Action Recognition. Neuroscience Applied.
American Psychiatric Association. (2023). Diagnostic and Statistical Manual of Mental Disorders (5th ed.).
Ginsberg, Y., Quintero, J., Anand, E., Casillas, M., & Upadhyaya, H. P. (2014). Underdiagnosis of attention-deficit/hyperactivity disorder in adults: A review. The Primary Care Companion for CNS Disorders, 16(3).