Taiwan study finds brain activity pattern reveals early signs of internet addiction

台灣一項研究發現，大腦活動模式有望揭示網路成癮的早期徵兆。

TAIPEI (Taiwan News) — Taiwanese researchers have developed an AI-based brainwave screening method to detect early signs of internet addiction, offering a more objective alternative to self-assessments.

The research team, led by National Health Research Institutes investigator Huang Hsu-wen (黃緒文), combined electroencephalography (EEG) with machine learning to analyze brain activity patterns associated with problematic internet use. The method achieved an accuracy rate of 86%, according to CNA.

Speaking at a press briefing Thursday, Huang described internet addiction as excessive and uncontrollable use of the internet or computers, often involving online gaming, social media, or pornography. The condition is most common among young adults aged 18 to 39.

Huang noted that internet addiction is linked to higher risks of depression and anxiety, as well as negative effects on academic performance and social functioning. It can also contribute to sleep deprivation and obesity, highlighting the importance of early detection.

For the study, researchers recruited 92 university students, including 42 already identified as having internet addiction and 50 without. While participants were awake and relaxed, resting-state brainwave data were collected, and the team analyzed connectivity between different brain regions.

Results showed that participants with internet addiction had significantly higher brain connectivity than students without. This was observed in the frontal lobe’s delta frequency band and across the whole brain in the gamma frequency band.

Researchers said the increased connectivity reflects stronger synchronized activity in regions responsible for attention, inhibitory control, and visual processing. Huang explained that these patterns are likely linked to impaired inhibitory control and heightened sensitivity to visual stimuli.

Together, these changes produce what the team described as an “excessive synchronization” of brain activity. Using these neural features as markers, the researchers applied multiple machine learning models to classify brainwave patterns.

Huang said these brainwave changes may appear before clear behavioral symptoms emerge. Conducting the test could allow for earlier detection, timely intervention, and more targeted support from schools and healthcare providers.