Creating personalized digital twins of the human brain and body excites neuroscientists and medical experts. These computer simulations mimic interactions among brain regions and predict responses to stimulation, diseases, or drugs. The brain’s billions of neurons pose immense challenges, even with AI and vast datasets. Existing whole-brain models often fail to reflect individual uniqueness, as each person’s neural connections form a distinct “brain fingerprint.”
Limitations of Current Models
Most digital brain twins resemble generic templates rather than true personal replicas. They perform no better than using a stranger’s wiring diagram. This gap undermines their value for simulating treatments before real-world application. Inaccurate models yield non-personalized predictions, potentially misleading clinicians in critical cases.
The Role of Competition in Brain Dynamics
The brain constantly shifts activity, trackable via neuroimaging like functional MRI. Personalized models simulate region interactions, but many overlook competition. While brains cooperate within circuits, regions vie for resources during attention shifts or multitasking. Over 20 years, simulations largely enforced cooperation, leading to unrealistic synchronized states.
Key Findings from Cross-Species Analysis
A comprehensive study across humans, macaque monkeys, and mice used brain activity recordings to compare models. Cooperative-only versions fell short, while those allowing excitation or suppression excelled. Competitive models better matched real cognitive circuits for attention and memory, based on over 14,000 neuroimaging datasets.
Competition stabilizes activity, enabling flexible region activation essential for intelligence. It prevents interference, boosting the brain’s energy efficiency far beyond AI systems.
Visual Insight
Models permitting competition generate patterns mirroring cognitive processes in humans, macaques, and mice.
Personalization and Broader Impacts
Competitive models capture unique brain fingerprints more accurately, enhancing individual specificity. These principles span species, aiding translational neuroscience. Animal trials often fail in humans—90% of neuropsychiatric treatments do not translate. Integrated human imaging and modeling could bridge this gap.
For conditions like epilepsy or tumors, digital twins simulate drug or stimulation effects, refining treatments beyond trial-and-error. Such frameworks also guide efficient AI development, paving the way for brain-faithful simulations and mind-like systems.
