
X, the Alphabet subsidiary trying to use artificial intelligence to discover a biomarker for depression and anxiety in brainwave data, reported this week that its initiative had failed to home in on a single indicator that could make measuring mental health as simple as reading a glucose meter.
Given the mystery that still shrouds the human brain, this was probably a predictable outcome for the initiative, dubbed Project Amber. But it stands as another reminder that machine learning is not magic pixie dust that can solve every problem in medicine.
Still, the company did find that advanced machine learning approaches were helpful in picking out features of electroencephalography (EEG) data that pointed to diagnoses such as major depressive disorder and generalized anxiety disorder.