The brain is an orchestra where each group of instruments must come in on time. In Alzheimer's disease, the musicians lose their notes: connections between brain regions get tangled, and the harmonious performance turns into cacophony. Scientists have long seen these disruptions on scans, but they didn't know if the situation could be fixed by touching just one string.

The researchers built a digital twin for each patient — a computer model that precisely replicates the pattern of disrupted connections in their brain. Initially, the model could identify the disease from this pattern, though not as well as from brain volume loss. But the authors were interested in something else: dynamics can be altered by stimulation, whereas lost tissue cannot. When the virtual patient's connections were 'corrected' to match a healthy pattern, the diagnosis disappeared. However, this required editing many points at once, like tuning the entire orchestra at the same time.

So the scientists took a different approach: for each person, they calculated the single region that most strongly influenced the model's final 'diagnosis.' A targeted strike on this area completely eliminated the signs of the disease. Moreover, a smart controller working in real time achieved the same result with a lower stimulation dose, relying only on available data. The ideal targets varied from person to person and were located in the cortex — not where connectivity was most disrupted, but where the network was most responsive to treatment. Thus, the conductor's baton restores the orchestra's lost harmony.