There is both clinical and neuroanatomical variability at the single-subject level in Alzheimer’s disease, complicating our understanding of brain-behaviour relationships and making it challenging to develop neuroimaging biomarkers to track disease severity, progression, and response to treatment. Prior work has shown that both group-level atrophy in clinical dementia syndromes and complex neurological symptoms in patients with focal brain lesions localize to brain networks. Here, we use a new technique termed ‘atrophy network mapping’ to test the hypothesis that single-subject atrophy maps in patients with a clinical diagnosis of Alzheimer’s disease will also localize to syndrome-specific and symptom-specific brain networks. First, we defined single-subject atrophy maps by comparing cortical thickness in each Alzheimer’s disease patient versus a group of age-matched, cognitively normal subjects across two independent datasets (total Alzheimer’s disease patients = 330). No more than 42% of Alzheimer’s disease patients had atrophy at any given location across these datasets. Next, we determined the network of brain regions functionally connected to each Alzheimer’s disease patient’s location of atrophy using seed-based functional connectivity in a large (n = 1000) normative connectome. Despite the heterogeneity of atrophied regions at the single-subject level, we found that 100% of patients with a clinical diagnosis of Alzheimer’s disease had atrophy functionally connected to the same brain regions in the mesial temporal lobe, precuneus cortex, and angular gyrus. Results were specific versus control subjects and replicated across two independent datasets. Finally, we used atrophy network mapping to define symptom-specific networks for impaired memory and delusions, finding that our results matched symptom networks derived from patients with focal brain lesions. Our study supports atrophy network mapping as a method to localize clinical, cognitive, and neuropsychiatric symptoms to brain networks, providing insight into brain-behaviour relationships in patients with dementia.

for the 4 Repeat Tau Neuroimaging Initiative † Objective: Perirolandic atrophy occurs in corticobasal syndrome (CBS) but is not specific versus progressive supranuclear palsy (PSP). There is heterogeneity in the locations of atrophy outside the perirolandic cortex and it remains unknown why atrophy in different locations would cause the same CBS-specific symptoms. In prior work, we used a wiring diagram of the brain called the human connectome to localize lesion-induced disorders to symptom-specific brain networks. Here, we use a similar technique termed "atrophy network mapping" to localize single-subject atrophy maps to symptom-specific brain networks. Methods: Single-subject atrophy maps were generated by comparing cortical thickness in patients with CBS versus controls. Next, we performed seed-based functional connectivity using a large normative connectome to determine brain regions functionally connected to each patient's atrophied locations. Results: Patients with CBS had perirolandic atrophy versus controls at the group level, but locations of atrophy in CBS were heterogeneous outside of the perirolandic cortex at the single-subject level (mean spatial correlation = 0.04). In contrast, atrophy occurred in locations functionally connected to the perirolandic cortex in all patients with CBS (spatial correlation = 0.66). Compared with PSP, patients with CBS had atrophy connected to a network of higher-order sensorimotor regions beyond perirolandic cortex, matching a CBS atrophy network from a recent meta-analysis. Finally, atrophy network mapping identified a symptom-specific network for alien limb, matching a lesion-induced alien limb network and a network associated with agency in healthy subjects. Interpretation: We identified a syndrome-specific network for CBS and symptom-specific network for alien limb using single-subject atrophy maps and the human connectome.

Emerging evidence has implicated the cerebellum in not just motor, but cognitive and behavioral roles in the brain. Previous research in cerebellar diseases have reported impulsive behaviors, but it is unknown if Essential Tremor (ET) patients present with these behaviors and where they localize. Here, in 105 ET patients, we measured disinhibition using the Frontal Systems Behavior Scale (FrSBe) and assessed the relationship between disinhibition scores and grey matter volume in the cerebellum, as well as functional correlates of disinhibition-related grey matter atrophy. These results are the first to evaluate neurobiological changes associated with disinhibition in ET.