Explore the vast range of titles available.

Brain lesions do not just disable but also disconnect brain areas, which once deprived of their input or output, can no longer subserve behaviour and cognition. The role of white matter connections has remained an open question for the past 250 years. Based on 1333 stroke lesions, here we reveal the human Disconnectome and demonstrate its relationship to the functional segregation of the human brain. Results indicate that functional territories are not only defined by white matter connections, but also by the highly stereotyped spatial distribution of brain disconnections. While the former has granted us the possibility to map 590 functions on the white matter of the whole brain, the latter compels a revision of the taxonomy of brain functions. Overall, our freely available Atlas of White Matter Function will enable improved clinical-neuroanatomical predictions for brain lesion studies and provide a platform for explorations in the domain of cognition. Brain disconnection after stroke leads to functional deficits whose anatomical basis is poorly understood. Here, based on a collection of stroke imaging, a database of neuroimaging meta-analysis, and high fidelity white matter mapping, the authors provide an atlas of human white matter function.

Key Points The dorsomedial frontal cortex contains a cluster of areas that are designated the supplementary motor area (SMA), the supplementary eye field (SEF) and the pre-supplementary motor area (pre-SMA). The defining functional feature of the members of this supplementary motor complex (SMC) is a marked sensitivity to various aspects of action. The anatomical features of the SMC are not homogeneous: there is a gradient of morphological and connectional change where affinity with the prefrontal and primary motor cortices changes reciprocally in the rostro–caudal plane. More-rostral regions show greater kinship with the prefrontal cortex than with the primary motor cortex; for more-caudal regions the reverse is true. The SMC is also heterogeneous neurophysiologically. The subregions of the SMC show different patterns of effector predilection and exhibit relative differences in the preponderance of cells that are sensitive to more-complex aspects of action. Compared with primary motor areas, the SMC exhibits greater sensitivity to tasks in which action contingencies are broader in range and not unambiguously specified by the immediate external environment. This contrast is illustrated by differences in SMC activity between 'self-initiated' and 'externally triggered' actions, by movement sequences, by well-learnt and poorly learnt actions, and by switching between action possibilities. Damage to the SMC disrupts behaviour in complex ways, affecting not just the commission but also the omission of actions, and the effects are broadly reflective of the neurophysiological properties of the region. This plurality of functional responses has traditionally been taken as implying a plurality of neural functions, including implementing intentions, learning and performing temporally organized actions, switching between actions, and inhibiting unwanted actions. Traditional accounts of the role of the SMC in voluntary action assume a fundamentally discrete modular architecture, with different functions assigned to macroscopically defined and functionally homogeneous regions. Thus, one function could be assigned to the SMA (for example, performing sequences) and another could be assigned to the pre-SMA (for example, changing between sequences). To explain the full range of behaviours, these discrete units must be able to switch between different functions depending on task demands — a feature we term functional pleomorphism. Functional pleomorphism is conceptually problematic owing to the difficulty of explaining the process of switching between different neural functions. Moreover, a discrete modular architecture implies greater functional and structural homogeneity within functional modules than between them, a premise that is not supported by the empirical data. The differences across the region are better described by smoothly varying continuities than by any kind of discrete pattern, and they therefore need correspondingly smooth models of functional organization. The plurality of functions that has been assigned to the SMC is undercut by a single elemental confound: the conditional complexity of the underlying condition–action association. Framed in the terms of information theory, internal, sequential, new and otherwise-flexible actions require more information than externally triggered, non-sequential and well-learnt actions (which were previously considered to be matched in complexity). We suggest that gaining further insight into the role of the SMC requires a unified account of the region that is based on a faithful picture of the anatomical and neurophysiological features and a conceptually rigorous analysis of the models supposed to explain them. The supplementary motor complex has a role in regulating action, but whether each of its subregions has a distinct function is unclear. Husain and colleagues review the literature and discuss outstanding issues regarding the function of this complex. The supplementary motor complex consists of the supplementary motor area, the supplementary eye field and the pre-supplementary motor area. In recent years, these areas have come under increasing scrutiny from cognitive neuroscientists, motor physiologists and clinicians because they seem to be crucial for linking cognition to action. However, theories regarding their function vary widely. This Review brings together the data regarding the supplementary motor regions, highlighting outstanding issues and providing new perspectives for understanding their functions.

Failure to attend scheduled hospital appointments disrupts clinical management and consumes resource estimated at £1 billion annually in the United Kingdom National Health Service alone. Accurate stratification of absence risk can maximize the yield of preventative interventions. The wide multiplicity of potential causes, and the poor performance of systems based on simple, linear, low-dimensional models, suggests complex predictive models of attendance are needed. Here, we quantify the effect of using complex, non-linear, high-dimensional models enabled by machine learning. Models systematically varying in complexity based on logistic regression, support vector machines, random forests, AdaBoost, or gradient boosting machines were trained and evaluated on an unselected set of 22,318 consecutive scheduled magnetic resonance imaging appointments at two UCL hospitals. High-dimensional Gradient Boosting Machine-based models achieved the best performance reported in the literature, exhibiting an area under the receiver operating characteristic curve of 0.852 and average precision of 0.511. Optimal predictive performance required 81 variables. Simulations showed net potential benefit across a wide range of attendance characteristics, peaking at £3.15 per appointment at current prevalence and call efficiency. Optimal attendance prediction requires more complex models than have hitherto been applied in the field, reflecting the complex interplay of patient, environmental, and operational causal factors. Far from an exotic luxury, high-dimensional models based on machine learning are likely essential to optimal scheduling amongst other operational aspects of hospital care. High predictive performance is achievable with data from a single institution, obviating the need for aggregating large-scale sensitive data across governance boundaries.

Magnetic resonance imaging and X-ray computed tomography provide the two principal methods available for imaging the brain at high spatial resolution, but these methods are not easily portable and cannot be applied safely to all patients. Ultrasound imaging is portable and universally safe, but existing modalities cannot image usefully inside the adult human skull. We use in silico simulations to demonstrate that full-waveform inversion, a computational technique originally developed in geophysics, is able to generate accurate three-dimensional images of the brain with sub-millimetre resolution. This approach overcomes the familiar problems of conventional ultrasound neuroimaging by using the following: transcranial ultrasound that is not obscured by strong reflections from the skull, low frequencies that are readily transmitted with good signal-to-noise ratio, an accurate wave equation that properly accounts for the physics of wave propagation, and adaptive waveform inversion that is able to create an accurate model of the skull that then compensates properly for wavefront distortion. Laboratory ultrasound data, using ex vivo human skulls and in vivo transcranial signals, demonstrate that our computational experiments mimic the penetration and signal-to-noise ratios expected in clinical applications. This form of non-invasive neuroimaging has the potential for the rapid diagnosis of stroke and head trauma, and for the provision of routine monitoring of a wide range of neurological conditions.

The presence of bias in artificial intelligence has garnered increased attention, with inequities in algorithmic performance being exposed across the fields of criminal justice, education, and welfare services. In health care, the inequitable performance of algorithms across demographic groups may widen health inequalities. Here, we identify and characterize bias in cardiology algorithms, looking specifically at algorithms used in the management of heart failure. Stage 1 involved a literature search of PubMed and Web of Science for key terms relating to cardiac machine learning (ML) algorithms. Papers that built ML models to predict cardiac disease were evaluated for their focus on demographic bias in model performance, and open-source data sets were retained for our investigation. Two open-source data sets were identified: (1) the University of California Irvine Heart Failure data set and (2) the University of California Irvine Coronary Artery Disease data set. We reproduced existing algorithms that have been reported for these data sets, tested them for sex biases in algorithm performance, and assessed a range of remediation techniques for their efficacy in reducing inequities. Particular attention was paid to the false negative rate (FNR), due to the clinical significance of underdiagnosis and missed opportunities for treatment. In stage 1, our literature search returned 127 papers, with 60 meeting the criteria for a full review and only 3 papers highlighting sex differences in algorithm performance. In the papers that reported sex, there was a consistent underrepresentation of female patients in the data sets. No papers investigated racial or ethnic differences. In stage 2, we reproduced algorithms reported in the literature, achieving mean accuracies of 84.24% (SD 3.51%) for data set 1 and 85.72% (SD 1.75%) for data set 2 (random forest models). For data set 1, the FNR was significantly higher for female patients in 13 out of 16 experiments, meeting the threshold of statistical significance (-17.81% to -3.37%; P<.05). A smaller disparity in the false positive rate was significant for male patients in 13 out of 16 experiments (-0.48% to +9.77%; P<.05). We observed an overprediction of disease for male patients (higher false positive rate) and an underprediction of disease for female patients (higher FNR). Sex differences in feature importance suggest that feature selection needs to be demographically tailored. Our research exposes a significant gap in cardiac ML research, highlighting that the underperformance of algorithms for female patients has been overlooked in the published literature. Our study quantifies sex disparities in algorithmic performance and explores several sources of bias. We found an underrepresentation of female patients in the data sets used to train algorithms, identified sex biases in model error rates, and demonstrated that a series of remediation techniques were unable to address the inequities present.

How do we best constrain social interactions to decrease transmission of communicable diseases? Indiscriminate suppression is unsustainable long term and presupposes that all interactions carry equal importance. Instead, transmission within a social network has been shown to be determined by its topology. In this paper, we deploy simulations to understand and quantify the impact on disease transmission of a set of topological network features, building a dataset of 9000 interaction graphs using generators of different types of synthetic social networks. Independently of the topology of the network, we maintain constant the total volume of social interactions in our simulations, to show how even with the same social contact some network structures are more or less resilient to the spread. We find a suitable intervention to be specific suppression of unfamiliar and casual interactions that contribute to the network’s global efficiency. This is, pathogen spread is significantly reduced by limiting specific kinds of contact rather than their global number. Our numerical studies might inspire further investigation in connection to public health, as an integrative framework to craft and evaluate social interventions in communicable diseases with different social graphs or as a highlight of network metrics that should be captured in social studies.

Distinctive patterns of brain neurotransmission frame determinant circuits for behavior. Understanding the relationship between their damage and the cognitive impairment provoked by brain lesions could provide insights into the pathophysiology and therapeutics of disabling disorders, like stroke. Yet, the challenges of neurotransmitter circuits mapping in vivo have hampered this investigation. Here, we developed an MRI white matter atlas of neurotransmitter circuits and created a method to chart how stroke damages neurotransmitter systems, which distinguishes pre and postsynaptic disruption. Our model, trained and tested in two large stroke patient samples, identified eight clusters with different neurochemical patterns. The associations with patients’ cognitive profiles were scarce, denoting that a particular cognitive deficit might have finer underlying neurochemical disturbances that are unfit to the granularity of our analyses. These findings depict stroke neurochemical diaschisis patterns, provide insights into stroke cognitive deficits and potential treatments, and open a new window for tailored neurotransmitter modulation. Neurotransmitters play a key role in brain function. Here, an MRI atlas of neurotransmitter circuits was developed to map stroke pre- and postsynaptic disruptions.

Characterizing acute service demand is critical for neurosurgery and other emergency-dominant specialties in order to dynamically distribute resources and ensure timely access to treatment. This is especially important in the post-Covid 19 pandemic period, when healthcare centers are grappling with a record backlog of pending surgical procedures and rising acute referral numbers. Healthcare dashboards are well-placed to analyze this data, making key information about service and clinical outcomes available to staff in an easy-to-understand format. However, they typically provide insights based on inference rather than prediction, limiting their operational utility. We retrospectively analyzed and prospectively forecasted acute neurosurgical referrals, based on 10,033 referrals made to a large volume tertiary neurosciences center in London, U.K., from the start of the Covid-19 pandemic lockdown period until October 2021 through the use of a novel AI-enabled predictive dashboard. As anticipated, weekly referral volumes significantly increased during this period, largely owing to an increase in spinal referrals (p < 0.05). Applying validated time-series forecasting methods, we found that referrals were projected to increase beyond this time-point, with Prophet demonstrating the best test and computational performance. Using a mixed-methods approach, we determined that a dashboard approach was usable, feasible, and acceptable among key stakeholders.

Chronic constipation is classified into 2 main syndromes, irritable bowel syndrome with constipation (IBS-C) and functional constipation (FC), on the assumption that they differ along multiple clinical characteristics and are plausibly of distinct pathophysiology. Our aim was to test this assumption by applying machine learning to a large prospective cohort of comprehensively phenotyped patients with constipation. Demographics, validated symptom and quality of life questionnaires, clinical examination findings, stool transit, and diagnosis were collected in 768 patients with chronic constipation from a tertiary center. We used machine learning to compare the accuracy of diagnostic models for IBS-C and FC based on single differentiating features such as abdominal pain (a \"unisymptomatic\" model) vs multiple features encompassing a range of symptoms, examination findings and investigations (a \"syndromic\" model) to assess the grounds for the syndromic segregation of IBS-C and FC in a statistically formalized way. Unisymptomatic models of abdominal pain distinguished between IBS-C and FC cohorts near perfectly (area under the curve 0.97). Syndromic models did not significantly increase diagnostic accuracy (P > 0.15). Furthermore, syndromic models from which abdominal pain was omitted performed at chance-level (area under the curve 0.56). Statistical clustering of clinical characteristics showed no structure relatable to diagnosis, but a syndromic segregation of 18 features differentiating patients by impact of constipation on daily life. IBS-C and FC differ only about the presence of abdominal pain, arguably a self-fulfilling difference given that abdominal pain inherently distinguishes the 2 in current diagnostic criteria. This suggests that they are not distinct syndromes but a single syndrome varying along one clinical dimension. An alternative syndromic segregation is identified, which needs evaluation in community-based cohorts. These results have implications for patient recruitment into clinical trials, future disease classifications, and management guidelines.

The gold standard in the treatment of ischaemic stroke is set by evidence from randomized controlled trials, typically using simple estimands of presumptively homogeneous populations. Yet the manifest complexity of the brain’s functional, connective, and vascular architectures introduces heterogeneities that violate the underlying statistical premisses, potentially leading to substantial errors at both individual and population levels. The counterfactual nature of interventional inference renders quantifying the impact of this defect difficult. Here we conduct a comprehensive series of semi-synthetic, biologically plausible, virtual interventional trials across 100M+ distinct simulations. We generate empirically grounded virtual trial data from large-scale meta-analytic connective, functional, genetic expression, and receptor distribution data, with high-resolution maps of 4K+ acute ischaemic lesions. Within each trial, we estimate treatment effects using models varying in complexity, in the presence of increasingly confounded outcomes and noisy treatment responses. Individualized prescriptions inferred from simple models, fitted to unconfounded data, are less accurate than those from complex models, even when fitted to confounded data. Our results indicate that complex modelling with richly represented lesion data may substantively enhance individualized prescriptive inference in ischaemic stroke. Stroke affects the brain in complex, highly individual ways. Here, the authors show that applying generative and causal AI methods to routinely collected brain scans may enable more closely personalized treatment recommendations.