Brain regions specialized for language have been extensively described, yet their arrangement into one or multiple networks remains debated. Using precision functional mapping across three independent cohorts of intensively scanned individuals (22 individuals scanned over 216 separate MRI sessions), we dissociated two nested left-lateralized perisylvian networks: an intermediate language network (intLANG) and an anatomically distinct association language network (aLANG). intLANG is anchored to precentral speech areas and the Sylvian parietal-temporal area (Spt), whereas aLANG surrounds intLANG and extends into higher-order prefrontal and temporal association cortices. The two networks can be fully recapitulated by functional connectivity from adjacent cerebellar regions, indicating that they are segregated, brain-wide networks. Task-based analyses further reveal that intLANG and aLANG are functionally distinct: intLANG responds robustly during rhyme judgments and nonword reading that emphasize phonology, whereas aLANG is preferentially recruited during meaning-based sentence processing. These findings indicate that human language engages nested distributed networks each specialized for distinct components of language processing: a lower-order network biased toward phonology, and a surrounding association network that subserves higher-order syntax and semantics. This nested organization is similar to other brain systems suggesting a shared hierarchical motif that may give rise to specialized cognitive functions across the human brain.

BACKGROUND AND OBJECTIVES:Early-onset Alzheimer disease (EOAD) is associated with substantial variability in clinical progression, and reliable biomarkers to predict the transition from mild cognitive impairment (MCI) to dementia remain limited. Structural MRI measures have demonstrated prognostic value in late-onset Alzheimer disease, but their utility for predicting progression in EOAD is less well understood. The goal was to examine whether baseline cortical atrophy predicts progression to dementia in patients with MCI because of EOAD. METHODS:This study included a well-characterized cohort of patients with EOAD enrolled in the large multisite natural history Longitudinal Early-Onset Alzheimer's Disease Study. Participants underwent standardized clinical assessments and structural MRI at baseline. Participants were aged between 40 and 64 years with biomarker-supported sporadic EOAD at the MCI stage. Cortical atrophy was measured within the EOAD-signature, a set of predominantly parieto-temporal regions showing greater atrophy in EOAD than in controls. Clinical severity was measured with the global Clinical Dementia Rating. Cox proportional hazards models estimated the association between baseline EOAD-signature atrophy burden and the hazard of progression to dementia over time. We evaluated whether EOAD-signature atrophy improved prognostic performance beyond baseline clinical severity using likelihood ratio tests, Akaike Information Criterion (AIC), and Harrell concordance index. RESULTS:A total of 130 patients with MCI due to EOAD (mean age 59.6 ± 4.1 years; 49% female) and 97 cognitively normal controls (mean age 56.9 ± 6.0 years; 64% female) were included. Greater baseline atrophy within the EOAD-signature predicted faster progression to dementia (hazard ratio [HR] = 1.24 per 1-SD increase in atrophy; 95% CI 1.13-1.37; p < 0.002). Adding EOAD-signature atrophy burden to a model including baseline clinical severity significantly improved model fit (ΔAIC = -4.5; likelihood ratio test p = 0.011). DISCUSSION:Baseline cortical atrophy within the EOAD-signature predicts progression from MCI to dementia in EOAD and provides prognostic information beyond baseline clinical severity. These findings support the potential value of EOAD-signature atrophy as an MRI-based biomarker for individualized prognostication and clinical trial stratification.

Longitudinal studies are required to measure individual differences in human brain aging, but are challenging over short intervals due to measurement error. Using cluster scanning, an approach that reduces error by densely repeating rapid structural scans, we assess brain aging in individuals across three timepoints in one year. Cluster scanning substantially improves the precision of individualized estimates, revealing previously undetectable individual differences in brain change. In just one year, we detect expected differences in the rates of brain aging between younger and older individuals, as well as differences between cognitively unimpaired and impaired individuals. Cognitively unimpaired older individuals variably reveal relative brain maintenance, unexpectedly rapid decline, and asymmetrical changes. We observe these atypical brain aging trajectories across structures and verify them in independent within-individual test-retest data. Cluster scanning promises to advance our understanding of the marked heterogeneity in brain aging by affording better short-term tracking of individual variability in structural change.

Regional brain atrophy estimated from structural magnetic resonance imaging (MRI) is a widely used measure of neurodegeneration in Alzheimer's disease (AD), Frontotemporal Lobar Degeneration (FTLD), and other dementias. Yet, traditional MRI-derived morphometric estimates are susceptible to measurement errors, posing a challenge for detecting longitudinal atrophy over short intervals. Here, we examined the utility of multiple MRI scans acquired in rapid succession (i.e., cluster scanning) for detecting longitudinal cortical atrophy over 3- and 6-month intervals within individual participants. Four individuals with mild cognitive impairment or mild dementia likely due to AD or FTLD participated in this study. At baseline, 3 months, and 6 months, structural MRI data were collected on a 3 Tesla scanner using a fast 1.2-mm T1-weighted multi-echo magnetization-prepared rapid gradient echo (MEMPRAGE) sequence (acquisition time = 2'23"). At each timepoint, participants underwent up to 32 MEMPRAGE scans acquired in four separate sessions over 2 days. Using linear mixed-effects models, we found that phenotypically vulnerable cortical ("core atrophy") regions exhibited statistically significant longitudinal atrophy in all participants (i.e., decreased cortical thickness) by 3 months and further demonstrated preferential vulnerability compared to control regions in three of the participants over at least one of the 3-month intervals. These findings provide proof-of-concept evidence that pooling multiple morphometric estimates derived from cluster scanning can detect longitudinal cortical atrophy over short intervals in individual patients with neurodegenerative dementias.

INTRODUCTION:Although cerebrospinal fluid (CSF) biomarkers reflect neurodegeneration in Alzheimer's disease (AD), it remains unclear whether these biomarkers track neurodegeneration in early-onset Alzheimer's disease (EOAD). METHODS:In 80 EOAD patients, we examined correlations between eight CSF biomarker levels and cortical thickness decreases within the EOAD cortical atrophy signature. Significant correlations were then entered into a multiple regression analysis. We also examined contributions of EOAD atrophy and CSF biomarkers to cognitive impairment. RESULTS:The levels of four CSF biomarkers correlated to EOAD signature atrophy. Stepwise regression analyses revealed that CSF neurofilament light chain (NfL) levels best predicted EOAD signature atrophy. Multiple regression showed EOAD signature atrophy combined with CSF NfL levels explained more variance in cognitive impairment than either factor alone. DISCUSSION:Within EOAD patients, CSF NfL levels relate to the magnitude of cortical atrophy, and a combination of EOAD signature atrophy and CSF NfL levels most robustly predict cognitive impairment.

Healthcare systems in the United States are now mandated to provide patients with immediate access to medical results. Access to complex results prior to appropriate clinical follow-up may cause confusion and fear, especially in populations with limited health literacy and English language proficiency. In this study, we demonstrate a use case of large language models (LLMs) to simplify and translate complex medical information for patients. We collected a consecutive series of brain MRI (N = 200), head CT (N = 100), and spine MRI (N = 200) reports, prompting a GPT-4o model to simplify these reports to a middle-school reading level and translate a subset of them to Spanish. Critically, two independent neuroradiology raters compared the simplified report impressions to the original report impressions, and found low rates of hallucinations, omissions, and imprecisions. The reading grade of the report impressions decreased from approximately college to approximately middle-school level, and readability improved across several text complexity metrics. Additionally, simplified impressions were classified as more neutral and less fearful. Machine generated translations of simplified impressions were statistically indistinguishable from manual translations as assessed by certified interpreters. In sum, we present a scalable, patient-friendly approach to promote healthcare literacy and engagement.

Abstract INTRODUCTION Behavioral variant frontotemporal dementia (bvFTD) is typically characterized by progressive deficits across several socioemotional and cognitive functions. Clinical prognostication in bvFTD remains a challenge due to the considerable variability in the rate of clinical progression and symptomatology. METHODS In a sample of sporadic bvFTD patients (N = 72), we generated cortical thickness estimates from baseline magnetic resonance imaging scans, which were used to predict longitudinal change in Clinical Dementia Rating Sum of Boxes (CDR‐SB) scores adapted for frontotemporal lobar degeneration (FTLD) (mean follow‐up duration = 2.00 ± 1.28 years). RESULTS The magnitude of cortical atrophy in regions of the default mode network that were largely spatially distinct from those exhibiting baseline atrophy most prominently and reliably predicted the rate of subsequent clinical decline. DISCUSSION We propose a latent network vulnerability model of clinical progression in neurodegenerative disease, which, with additional empirical support, may provide a unified framework for estimating person‐specific trajectories of clinical decline. Highlights At baseline, bvFTD is associated with prominent atrophy in anterior cortical regions. Atrophy in the posterior default mode network predicts subsequent clinical decline. Network‐based atrophy has potential to be a useful clinical prognostication tool in bvFTD.

Prefrontal regions are hypothesized to be organized hierarchically in support of cognitive control. Using precision functional MRI in three independent cohorts of intensively scanned participants (N=37), we consistently identified a lateral prefrontal cortex (LPFC) region linked to a canonical control network separate from a nearby rostral LPFC region linked to the action-mode network. Despite their spatial juxtaposition, the two regions were differentially coupled to the caudate and ventral putamen, suggesting they are components of segregated networks. Working memory demands activated the canonical LPFC control region but not the adjacent region. Contrasting go and no-go trials during target detection revealed a robust functional double dissociation during goal-directed behavior. The canonical LPFC control region activated when responses were withheld, while the putamen-coupled LPFC region increased activity during executed responses. These findings demonstrate that adjacent LPFC regions participate in opposing functions predicted by their embedding within distinct parallel large-scale networks.

PPA refers to a clinical syndrome presenting with language impairment with relative preservation of other cognitive functions. Neuroimaging evidence suggests that the language network, anchored in left prefrontal and temporo-parietal cortices, is selectively affected in PPA. To date, no pharmacological or neuromodulation strategies can satisfactorily improve symptoms in PPA. Focal neuromodulation techniques, such as repetitive TMS (rTMS), can change resting-state functional connectivity in a network-specific manner. Thus, we investigated whether rTMS could selectively modulate functional connections within the degenerated language network in PPA. A double-blinded, sham controlled, cross-over design was used to administer intermittent theta burst stimulation (iTBS) for 10 days in a heterogeneous sample of PPA patients: 4 logopenic variant (lvPPA), 2 non-fluent variant (nfvPPA), 1 semantic variant (svPPA), and 3 with primary progressive apraxia of speech (PPAOS). We stimulated the left caudal middle frontal gyrus region most functionally correlated with the language network on an individual-subject basis. Standardized language and functional communication assessments were administered at baseline, post-active TMS, and post-sham TMS. Resting-state fMRI data were analyzed to probe functional connectivity changes across sessions. We found language improvements selective to active TMS in the functional communication analysis and standardized language assessment in all PPA subtypes and PPAOS. Across all subtypes, increased functional connectivity was observed in the language and other networks subserving cognitive performance after active TMS. We demonstrate preliminary evidence that personalized functional-network guided iTBS can improve language impairments in a heterogeneous PPA and PPAOS cohort. Furthermore, improvements in language tests were accompanied by increases in functional network connectivity, pointing to a putative neural mechanism of TMS-induced benefits in PPA and PPAOS.

Functional neuroimaging studies have described changes in local activity or in distributed connectivity in Major Depressive Disorder (MDD). We studied the interactions between these levels of neural organization to reveal novel pathobiological signatures of MDD. We scanned 21 unmedicated MDD subjects and 21 matched non-depressed controls at rest with simultaneous 18fluorodeoxyglucose PET (FDG-PET, to measure local metabolic activity), and resting-state functional Magnetic Resonance Imaging (rsfMRI, to measure hemodynamic functional connectivity, FC). We first identified regions of interest (ROIs) with increased (hypermetabolic) or decreased (hypometabolic) glucose metabolism in depressed compared to non-depressed subjects. Next, we used these hypermetabolic and hypometabolic regions as seed-based ROIs to derive FC maps, and compared these maps across depressed and non-depressed subjects. Depressed subjects demonstrated hypermetabolism in regions partially overlapping with the cingulo-opercular network, and hypometabolism in regions partially overlapping with the default network. In depressed subjects, hypermetabolic ROIs displayed stronger FC to portions of the frontostriatal salience network. Interestingly, hypometabolic ROIs displayed weaker FC with almost identical frontostriatal salience network regions. This work further emphasizes network-level metabolic and hemodynamic functional connectivity changes in MDD. More specifically, our findings suggest that local metabolic activity is up- or down-regulated in a distributed fashion in depression and that these metabolic changes are associated with comparable increases and decreases in functional coupling to the frontostriatal salience network-further bolstering the putative pathogenic importance of this network in MDD. Clinical Trials registration number: NCT01931995.

Higher-order cognitive and affective functions are supported by large-scale networks in the brain. Dysfunction in different networks is proposed to associate with distinct symptoms in neuropsychiatric disorders. However, the specific networks targeted by current clinical transcranial magnetic stimulation (TMS) approaches are unclear. While standard-of-care TMS relies on scalp-based landmarks, recent FDA-approved TMS protocols use individualized functional connectivity with the subgenual anterior cingulate cortex (sgACC) to optimize TMS targeting. Leveraging previous work on precision network estimation and modeling of the TMS electric field (E-field), we asked whether various clinical TMS approaches target different functional networks between individuals. Results revealed that modeled homotopic scalp positions (left F3 and right F4) target different networks within and across individuals, and right F4 generally favors a right-lateralized control network. TMS coil positions over the dorsolateral prefrontal cortex (dlPFC) zone anticorrelated with the sgACC most frequently target a network coupled to the ventral striatum (reward circuitry) but largely miss that network in some individuals. We further illustrate how modeling can be used to retrospectively assess the estimated targets achieved in prior TMS sessions and also used to prospectively provide coil positions that can target distinct closely localized dlPFC network regions with spatial selectivity and maximal E-field intensity. In a final study, precision targeting was found to be feasible in participants with Major Depressive Disorder using data derived from a single low-burden MRI session suggesting the methods are applicable to translational efforts where limiting patient burden and ensuring robustness are critical.

Behavioral variant frontotemporal dementia (bvFTD) is a clinically heterogeneous syndrome characterized by progressive deficits across a range of socioemotional and cognitive functions, such as apathy, disinhibition, loss of empathy, impulsivity, and executive dysfunction. Clinical prognostication in bvFTD remains a challenge due to the considerable variability in the rate of clinical progression and symptomatology, and the field currently lacks robust and reliable tools. Here, we investigated the utility of MRI-based cortical atrophy as a predictor of longitudinal clinical decline in a sample of sporadic bvFTD patients. We analyzed data obtained from two independent cohorts of bvFTD patients: the Massachusetts General Hospital cohort (n = 39) and the FTLD-NI cohort (n = 33). All patients had cortical thickness estimates from baseline MRI scans, which were used to predict longitudinal change in clinical impairment assessed by the CDR Sum-of-Boxes score adapted for FTLD (CDR FTLD-SB). Multivariate partial least squares (PLS) analysis was performed to identify a weighted linear combination of cortical vertices where the magnitude of atrophy is maximally associated with the rate of CDR FTLD-SB change across bvFTD patients. The current sample included 72 patients with sporadic bvFTD (Table 1). Each patient had at least one clinical follow-up, with the mean duration of 2.00 ± 1.28 years. On average, CDR FTLD-SB scores increased by 4.40 points annually (p < .0001) (Figure 1). bvFTD patients exhibited prominent cortical atrophy at baseline in anterior cortical regions, including those part of the limbic network (e.g., temporal pole) as well as the anterior nodes of the default mode network (e.g., dorsomedial prefrontal cortex, rostral lateral temporal cortex) (Figure 2A) Multivariate vertex-wise PLS analysis showed that default mode network regions spatially distinct from those exhibiting baseline atrophy (e.g., ventromedial prefrontal cortex, posteromedial cortex, inferior parietal lobule) most prominently and reliably predicted the rate of subsequent clinical decline (Figure 2B). These results point to the clinical prognostic utility of MRI-derived cortical atrophy in sporadic bvFTD. We propose a latent network vulnerability model of clinical progression in neurodegenerative disease, which may provide a unified framework for estimating person-specific trajectories of clinical decline and optimizing clinical trial designs.

The ventral striatum (VS) receives input from the cerebral cortex and is modulated by midbrain dopaminergic projections in support of processing reward and motivation. Here, we explored the organization of cortical regions linked to the human VS using within-individual functional connectivity MRI (fcMRI) in intensively scanned participants. In two initial participants (scanned 31 sessions each), seed regions in the VS were preferentially correlated with distributed cortical regions that are part of the salience network. The VS seed regions recapitulated salience network topography and replicated in each individual, including anterior and posterior midline regions, anterior insula, and dorsolateral prefrontal cortex (DLPFC). The topography was distinct from adjacent striatal seed regions and from cortical networks associated with domain-flexible cognitive control. Unbiased comprehensive analyses of the full striatum confirmed that the VS is coupled to the salience network while also revealing the established, spatially separated cognitive zones of the caudate and motor zones of the putamen. VS correlation with the salience network, including DLPFC, was observed in 15 additional participants (scanned 8 or more times each), indicating it is a robust and generalizable finding. These results suggest that the VS contributes to a cortico-basal ganglia loop that is part of the salience network and raise the possibility that the DLPFC may be an effective neuromodulatory target for neuropsychiatric disorders of reward and motivation because of its preferential coupling to the VS.NEW & NOTEWORTHY Individualized precision neuroimaging reveals the ventral striatum (VS) is preferentially correlated with the salience network, including a region in the dorsolateral prefrontal cortex (DLPFC) that is adjacent to regions associated with cognitive control. These results raise the possibility that DLPFC is an effective neuromodulatory target for depression due to preferential coupling with the VS.

Precision mapping of brain networks within individuals prevailingly relies on functional connectivity analysis of resting-state data. Here, we explored whether networks can be estimated using only task data. Correlation matrices estimated from task data were similar to those derived from resting-state data. The largest factor affecting similarity was the amount of data. Precision networks estimated from task data showed strong spatial overlap with those derived from resting-state data and predicted the same triple functional dissociation in independent data. To illustrate novel possibilities enabled by the present methods, we mapped the detailed organization of thalamic association zones within individuals by pooling extensive resting-state and task data. We also demonstrated how task data can be used to estimate networks while simultaneously extracting task responses. Broadly, these findings suggest that there is an underlying, stable network architecture that is idiosyncratic to the individual and persists across task states.

The hippocampus possesses anatomical differences along its long axis. Here, we explored the functional specialization of the human hippocampal long axis using network-anchored precision functional MRI in two independent datasets (N = 11 and N = 9) paired with behavioral analysis (N = 266 and N = 238). Functional connectivity analyses demonstrated that the anterior hippocampus was preferentially correlated with a cerebral network associated with remembering, while the posterior hippocampus selectively contained a region correlated with a distinct network associated with behavioral salience. Seed regions placed within the hippocampus recapitulated the distinct cerebral networks. Functional characterization of the anterior and posterior hippocampal regions using task data identified and replicated a functional double dissociation. The anterior hippocampal region was sensitive to remembering and imagining the future, specifically tracking the process of scene construction, while the posterior hippocampal region displayed transient responses to targets in an oddball detection task and to transitions between task blocks. These findings suggest an unexpected specialization along the long axis of the human hippocampus with differential responses reflecting the functional properties of the partner cerebral networks.

Posterior cortical atrophy (PCA) is a clinical syndrome characterized by visuospatial and visuoperceptual symptoms arising from neurodegeneration of the posterior temporoparietal and occipital cortices. Focal neuromodulation such as repetitive transcranial magnetic stimulation (rTMS) can change resting-state functional connectivity in a network-specific manner. To investigate whether intermittent theta burst stimulation (iTBS) (a form of rTMS), can selectively modulate resting-state functional connectivity (FC) in posterior nodes of the default mode network (DMN) in a patient with PCA with amnestic symptoms. A stimulation target within the left caudal middle frontal (cMFG) was defined based on the maximum FC with the DMN. This target was then stimulated with iTBS in a double-blind placebo-controlled paradigm. Changes in the FC of the stimulation target was used as the outcome measure. Following active stimulation, the patient exhibited selective increase in functional connectivity between the stimulation target and the DMN. This case report provides novel evidence that network rTMS targeting using individualized resting-state fMRI can increase network FC in a patient with PCA. Larger controlled trials are needed to determine the impact of neuromodulation on cognitive performance in PCA and to further explore changes in DMN functional connectivity as a mechanism of these effects.

Longitudinal studies are required to measure individual differences in human brain aging, but they are difficult to estimate over short intervals because of measurement error. Using cluster scanning, an approach that reduces error by densely repeating rapid structural scans, we assessed brain aging in individuals across three longitudinal timepoints spaced across one year. Cluster scanning substantially improved the precision of individualized estimates, revealing previously undetectable individual differences in brain change. In just one year, expected differences in the rates of brain aging between younger and older individuals were evident, as were differences between cognitively unimpaired and impaired individuals. Each person's brain change trajectory was compared to modeled normative expectations from a large cohort of age-matched UK Biobank participants. Cognitively unimpaired older individuals variably revealed relative brain maintenance, unexpectedly rapid decline, and asymmetrical changes. These atypical brain aging trajectories were found across structures and verified in independent within-individual test and retest data. Cluster scanning promises to advance our understanding of the marked heterogeneity in brain aging by affording better short-term tracking of individual variability in structural change.

There is a significant need for biomarkers of neurodegenerative burden in Early-onset Alzheimer’s disease (EOAD). Evidence suggests that levels of specific CSF biomarkers (e.g., Neurofilament light (NfL), Synaptosomal-Associated Protein (SNAP-25), neurogranin, Visinin-like protein 1 (VILIP-1), Aß 42/40, phospho-tau and total tau) index the extent of neurodegeneration in dementing illnesses. However, it remains unclear whether these biomarkers correlate to cortical atrophy patterns in EOAD. Based on prior work demonstrating correlations between NfL, SNAP-25, neurogranin and Aß 42/40 CSF levels and cognitive impairment in EOAD (Dage et al. 2023), we hypothesized that these biomarkers (and not VILIP-1, phospho-tau or total tau) would variably predict cortical atrophy within our recently described EOAD signature (Touroutoglou et al. 2023). We recruited 92 EOAD patients. In each patient, atrophy within the EOAD cortical signature were calculated as W-scores (i.e., Z-scores adjusted for age and sex relative to a sample of healthy controls). We first ran a simple regression analysis of each of the 7 CSF biomarkers and W-scores in the EOAD signature across EOAD patients. We then entered the biomarkers with significant correlations to the EOAD signature into a stepwise regression analysis with backward elimination to ascertain the most parsimonious model predicting atrophy in the EOAD signature. As a control region not expected to be related to these biomarkers, we assessed correlations to calcarine fissure cortical atrophy. As predicted, we observed a significant correlation between CSF levels of NfL, SNAP-25, neurogranin and Aß 42/40 and atrophy within the EOAD signature. After entering these four biomarkers into the stepwise regression analysis, the most parsimonious model identified complementary contributions of NfL, SNAP-25, and Aß 42/40, in predicting atrophy in the EOAD signature. There were no correlations between any biomarker and calcarine atrophy. Selected CSF biomarkers in EOAD patients predict the degree of atrophy within the EOAD cortical signature. Ongoing work includes correlating these biomarkers with topographical atrophy patterns on the whole-brain voxel-wise level. Our results suggest that certain CSF biomarkers could assess neurodegenerative burden within EOAD individuals. This would provide valuable information regarding disease progression for clinical care and clinical trials involving disease-modifying therapies.