Artificial intelligence for diagnostic and prognostic neuroimaging in dementia: A systematic review.

INTRODUCTION: Artificial intelligence (AI) and neuroimaging offer new opportunities for diagnosis and prognosis of dementia. METHODS: We systematically reviewed studies reporting AI for neuroimaging in diagnosis and/or prognosis of cognitive neurodegenerative diseases. RESULTS: A total of 255 studies were identified. Most studies relied on the Alzheimer's Disease Neuroimaging Initiative dataset. Algorithmic classifiers were the most commonly used AI method (48%) and discriminative models performed best for differentiating Alzheimer's disease from controls. The accuracy of algorithms varied with the patient cohort, imaging modalities, and stratifiers used. Few studies performed validation in an independent cohort. DISCUSSION: The literature has several methodological limitations including lack of sufficient algorithm development descriptions and standard definitions. We make recommendations to improve model validation including addressing key clinical questions, providing sufficient description of AI methods and validating findings in independent datasets. Collaborative approaches between experts in AI and medicine will help achieve the promising potential of AI tools in practice. HIGHLIGHTS: There has been a rapid expansion in the use of machine learning for diagnosis and prognosis in neurodegenerative disease Most studies (71%) relied on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset with no other individual dataset used more than five times There has been a recent rise in the use of more complex discriminative models (e.g., neural networks) that performed better than other classifiers for classification of AD vs healthy controls We make recommendations to address methodological considerations, addressing key clinical questions, and validation We also make recommendations for the field more broadly to standardize outcome measures, address gaps in the literature, and monitor sources of bias.

Tau burden and the functional connectome in Alzheimer's disease and progressive supranuclear palsy.

Alzheimer's disease and progressive supranuclear palsy (PSP) represent neurodegenerative tauopathies with predominantly cortical versus subcortical disease burden. In Alzheimer's disease, neuropathology and atrophy preferentially affect 'hub' brain regions that are densely connected. It was unclear whether hubs are differentially affected by neurodegeneration because they are more likely to receive pathological proteins that propagate trans-neuronally, in a prion-like manner, or whether they are selectively vulnerable due to a lack of local trophic factors, higher metabolic demands, or differential gene expression. We assessed the relationship between tau burden and brain functional connectivity, by combining in vivo PET imaging using the ligand AV-1451, and graph theoretic measures of resting state functional MRI in 17 patients with Alzheimer's disease, 17 patients with PSP, and 12 controls. Strongly connected nodes displayed more tau pathology in Alzheimer's disease, independently of intrinsic connectivity network, validating the predictions of theories of trans-neuronal spread but not supporting a role for metabolic demands or deficient trophic support in tau accumulation. This was not a compensatory phenomenon, as the functional consequence of increasing tau burden in Alzheimer's disease was a progressive weakening of the connectivity of these same nodes, reducing weighted degree and local efficiency and resulting in weaker 'small-world' properties. Conversely, in PSP, unlike in Alzheimer's disease, those nodes that accrued pathological tau were those that displayed graph metric properties associated with increased metabolic demand and a lack of trophic support rather than strong functional connectivity. Together, these findings go some way towards explaining why Alzheimer's disease affects large scale connectivity networks throughout cortex while neuropathology in PSP is concentrated in a small number of subcortical structures. Further, we demonstrate that in PSP increasing tau burden in midbrain and deep nuclei was associated with strengthened cortico-cortical functional connectivity. Disrupted cortico-subcortical and cortico-brainstem interactions meant that information transfer took less direct paths, passing through a larger number of cortical nodes, reducing closeness centrality and eigenvector centrality in PSP, while increasing weighted degree, clustering, betweenness centrality and local efficiency. Our results have wide-ranging implications, from the validation of models of tau trafficking in humans to understanding the relationship between regional tau burden and brain functional reorganization.

18F-AV-1451 positron emission tomography in Alzheimer's disease and progressive supranuclear palsy.

The ability to assess the distribution and extent of tau pathology in Alzheimer's disease and progressive supranuclear palsy in vivo would help to develop biomarkers for these tauopathies and clinical trials of disease-modifying therapies. New radioligands for positron emission tomography have generated considerable interest, and controversy, in their potential as tau biomarkers. We assessed the radiotracer 18F-AV-1451 with positron emission tomography imaging to compare the distribution and intensity of tau pathology in 15 patients with Alzheimer's pathology (including amyloid-positive mild cognitive impairment), 19 patients with progressive supranuclear palsy, and 13 age- and sex-matched controls. Regional analysis of variance and a support vector machine were used to compare and discriminate the clinical groups, respectively. We also examined the 18F-AV-1451 autoradiographic binding in post-mortem tissue from patients with Alzheimer's disease, progressive supranuclear palsy, and a control case to assess the 18F-AV-1451 binding specificity to Alzheimer's and non-Alzheimer's tau pathology. There was increased 18F-AV-1451 binding in multiple regions in living patients with Alzheimer's disease and progressive supranuclear palsy relative to controls [main effect of group, F(2,41) = 17.5, P < 0.0001; region of interest × group interaction, F(2,68) = 7.5, P < 0.00001]. More specifically, 18F-AV-1451 binding was significantly increased in patients with Alzheimer's disease, relative to patients with progressive supranuclear palsy and with control subjects, in the hippocampus and in occipital, parietal, temporal, and frontal cortices (t's > 2.2, P's < 0.04). Conversely, in patients with progressive supranuclear palsy, relative to patients with Alzheimer's disease, 18F-AV-1451 binding was elevated in the midbrain (t = 2.1, P < 0.04); while patients with progressive supranuclear palsy showed, relative to controls, increased 18F-AV-1451 uptake in the putamen, pallidum, thalamus, midbrain, and in the dentate nucleus of the cerebellum (t's > 2.7, P's < 0.02). The support vector machine assigned patients' diagnoses with 94% accuracy. The post-mortem autoradiographic data showed that 18F-AV-1451 strongly bound to Alzheimer-related tau pathology, but less specifically in progressive supranuclear palsy. 18F-AV-1451 binding to the basal ganglia was strong in all groups in vivo. Postmortem histochemical staining showed absence of neuromelanin-containing cells in the basal ganglia, indicating that off-target binding to neuromelanin is an insufficient explanation of 18F-AV-1451 positron emission tomography data in vivo, at least in the basal ganglia. Overall, we confirm the potential of 18F-AV-1451 as a heuristic biomarker, but caution is indicated in the neuropathological interpretation of its binding. Off-target binding may contribute to disease profiles of 18F-AV-1451 positron emission tomography, especially in primary tauopathies such as progressive supranuclear palsy. We suggest that 18F-AV-1451 positron emission tomography is a useful biomarker to assess tau pathology in Alzheimer's disease and to distinguish it from other tauopathies with distinct clinical and pathological characteristics such as progressive supranuclear palsy.

Incidence and distribution of injuries associated with e-scooter use: a radiological perspective.

OBJECTIVES: To investigate electric scooter (e-scooter)-associated radiological injury incidence and distribution of injuries. METHODS: Retrospective cross-sectional study of radiological examinations related to e-scooter injuries at a major trauma centre in a small university city. The hospital radiology information system was searched for terms related to e-scooters between January 1, 2015, and October 31, 2022. E-scooter use was confirmed by review of the patients' electronic medical records. Specific injuries were divided based on site of injury using the Injury Severity Scale categorized groups. RESULTS: A total of 568 radiological studies related to e-scooter injuries were identified on 340 distinct patients (56% male, with an average age of 28 years). Peak incidence of e-scooter-related injuries was seen in the summer months, after a local scooter sharing system was introduced in October 2020. A total of 149 patients had radiologically diagnosed injuries, with extremity injuries being most frequent (80%). Facial (8%), head/neck (8%), and thorax/abdomen (4%) were less common. Radial head fractures were the most common injury (n = 27). Thirteen patients had multiple sites of injury, four of which had both upper limb and facial bone fractures described. CONCLUSIONS: We report a significant increase in radiological investigations and injuries in the context of e-scooter injuries, particularly since the introduction of an e-scooter sharing scheme. This study informs radiologists on common locations of injuries when reporting studies of patients that have had e-scooter-related injuries. ADVANCES IN KNOWLEDGE: This is the first UK-based study providing a comprehensive radiological perspective of the impact of e-scooter use and associated distribution of injuries, adding important data for many cities that are currently undertaking review of their e-scooter sharing schemes.

Performance of ChatGPT on the Situational Judgement Test-A Professional Dilemmas-Based Examination for Doctors in the United Kingdom.

BACKGROUND: ChatGPT is a large language model that has performed well on professional examinations in the fields of medicine, law, and business. However, it is unclear how ChatGPT would perform on an examination assessing professionalism and situational judgement for doctors. OBJECTIVE: We evaluated the performance of ChatGPT on the Situational Judgement Test (SJT): a national examination taken by all final-year medical students in the United Kingdom. This examination is designed to assess attributes such as communication, teamwork, patient safety, prioritization skills, professionalism, and ethics. METHODS: All questions from the UK Foundation Programme Office's (UKFPO's) 2023 SJT practice examination were inputted into ChatGPT. For each question, ChatGPT's answers and rationales were recorded and assessed on the basis of the official UK Foundation Programme Office scoring template. Questions were categorized into domains of Good Medical Practice on the basis of the domains referenced in the rationales provided in the scoring sheet. Questions without clear domain links were screened by reviewers and assigned one or multiple domains. ChatGPT's overall performance, as well as its performance across the domains of Good Medical Practice, was evaluated. RESULTS: Overall, ChatGPT performed well, scoring 76% on the SJT but scoring full marks on only a few questions (9%), which may reflect possible flaws in ChatGPT's situational judgement or inconsistencies in the reasoning across questions (or both) in the examination itself. ChatGPT demonstrated consistent performance across the 4 outlined domains in Good Medical Practice for doctors. CONCLUSIONS: Further research is needed to understand the potential applications of large language models, such as ChatGPT, in medical education for standardizing questions and providing consistent rationales for examinations assessing professionalism and ethics.

A rare presentation of rapidly progressing myopathy in an adolescent.

We present a case of severe juvenile dermatomyositis with limited response to steroids in an adolescent who developed symptoms within hours after receiving Pfizer BNT162b2 coronavirus disease 2019 vaccine. The patient presented with severe weakness of proximal muscles, dyspnoea, and tachycardia. His muscle enzymes were raised, and he was diagnosed with severe juvenile dermatomyositis following magnetic resonance imaging and muscle biopsy. His management was challenging, requiring multidisciplinary input, and difficult decisions with regard to the appropriate immunomodulatory treatments. The patient had to undergo escalating immunosuppressive treatments before he began to recover clinically and biochemically. To our knowledge, this is the first case in an adolescent although a few cases of similar presentations following coronavirus disease 2019 vaccination have been reported in adults. Elucidating the potential relationship of the vaccine with this severe myopathy in an adolescent is important for global vaccination policies, but avoiding the conflation of association with causation is also crucial in the context of the pandemic.

P2Y12 inhibitors for the neurointerventionalist.

The use of antiplatelets is widespread in clinical practice. However, for neurointerventional procedures, protocols for antiplatelet use are scarce and practice varies between individuals and institutions. This is further complicated by the quantity of antiplatelet agents which differ in route of administration, dosage, onset of action, efficacy and ischemic and hemorrhagic complications. Clarifying the individual characteristics for each antiplatelet agent, and their associated risks, will increasingly become relevant as the practice of mechanical thrombectomy, stenting, coiling and flow diversion procedures grows. The aim of this review is to summarize the existing literature for the use of P2Y12 inhibitors in neurointerventional procedures, examine the quality of the evidence, and highlight areas in need of further research.

Glycoprotein IIb/IIIa inhibitors for the neurointerventionalist.

Antiplatelet therapies are commonly used in neurointerventional procedures. However, specific guidelines for their use in these settings is lacking and it can often be difficult to balance the potential risks and benefits of these medications. Considering the continued growth and adoption of neurointerventional procedures, it is crucial to understand the properties of these agents in order to use them safely. Large-scale clinical trials are still needed to clarify many of these aspects for this emerging field. However, the existing literature already provides insight into which antiplatelet drugs are of benefit to the neurointerventionalist as well as their associated risks of ischemic and hemorrhagic complications. Hence, this review focuses on the applications of GPIIb/IIIA inhibitors to neurointerventional procedures.

Atomoxetine and citalopram alter brain network organization in Parkinson's disease.

Parkinson's disease has multiple detrimental effects on motor and cognitive systems in the brain. In contrast to motor deficits, cognitive impairments in Parkinson's disease are usually not ameliorated, and can even be worsened, by dopaminergic treatments. Recent evidence has shown potential benefits from restoring other neurotransmitter deficits, including noradrenergic and serotonergic transmission. Here, we study global and regional brain network organization using task-free imaging (also known as resting-state), which minimizes performance confounds and the bias towards predetermined networks. Thirty-three patients with idiopathic Parkinson's disease were studied three times in a double-blinded, placebo-controlled counter-balanced crossover design, following placebo, 40 mg oral atomoxetine (selective noradrenaline reuptake inhibitor) or 30 mg oral citalopram (selective serotonin reuptake inhibitor). Neuropsychological assessments were performed outside the scanner. Seventy-six controls were scanned without medication to provide normative data for comparison to the patient cohort. Graph theoretical analysis of task-free brain connectivity, with a random 500-node parcellation, was used to measure the effect of disease in placebo-treated state (versus unmedicated controls) and pharmacological intervention (drug versus placebo). Relative to controls, patients on placebo had executive impairments (reduced fluency and inhibitory control), which was reflected in dysfunctional network dynamics in terms of reduced clustering coefficient, hub degree and hub centrality. In patients, atomoxetine improved fluency in proportion to plasma concentration (P = 0.006, r 2 = 0.24), and improved response inhibition in proportion to increased hub Eigen centrality (P = 0.044, r 2 = 0.14). Citalopram did not improve fluency or inhibitory control, but its influence on network integration and efficiency depended on disease severity: clustering (P = 0.01, r 2 = 0.22), modularity (P = 0.043, r 2 = 0.14) and path length (P = 0.006, r 2 = 0.25) increased in patients with milder forms of Parkinson's disease, but decreased in patients with more advanced disease (Unified Parkinson's Disease Rating Scale motor subscale part III > 30). This study supports the use of task-free imaging of brain networks in translational pharmacology of neurodegenerative disorders. We propose that hub connectivity contributes to cognitive performance in Parkinson's disease, and that noradrenergic treatment strategies can partially restore the neural systems supporting executive function.

[11C]PK11195 binding in Alzheimer disease and progressive supranuclear palsy.

OBJECTIVE: We tested whether in vivo neuroinflammation relates to the distinctive distributions of pathology in Alzheimer disease (AD) and progressive supranuclear palsy (PSP). METHODS: Sixteen patients with symptomatic AD (including amnestic mild cognitive impairment with amyloid-positive PET scan), 16 patients with PSP-Richardson syndrome, and 13 age-, sex-, and education-matched healthy controls were included in this case-control study. Participants underwent [11C]PK11195 PET scanning, which was used as an in vivo index of neuroinflammation. RESULTS: [11C]PK11195 binding in the medial temporal lobe and occipital, temporal, and parietal cortices was increased in patients with AD, relative both to patients with PSP and to controls. Compared to controls, patients with PSP showed elevated [11C]PK11195 binding in the thalamus, putamen, and pallidum. [11C]PK11195 binding in the cuneus/precuneus correlated with episodic memory impairment in AD, while [11C]PK11195 binding in the pallidum, midbrain, and pons correlated with disease severity in PSP. CONCLUSIONS: Together, our results suggest that neuroinflammation has an important pathogenic role in the 2 very different human neurodegenerative disorders of AD and PSP. The increase and distribution of microglial activation suggest that immunotherapeutic strategies may be useful in slowing the progression of both diseases.

Atomoxetine Enhances Connectivity of Prefrontal Networks in Parkinson's Disease.

Cognitive impairment is common in Parkinson's disease (PD), but often not improved by dopaminergic treatment. New treatment strategies targeting other neurotransmitter deficits are therefore of growing interest. Imaging the brain at rest ('task-free') provides the opportunity to examine the impact of a candidate drug on many of the brain networks that underpin cognition, while minimizing task-related performance confounds. We test this approach using atomoxetine, a selective noradrenaline reuptake inhibitor that modulates the prefrontal cortical activity and can facilitate some executive functions and response inhibition. Thirty-three patients with idiopathic PD underwent task-free fMRI. Patients were scanned twice in a double-blind, placebo-controlled crossover design, following either placebo or 40-mg oral atomoxetine. Seventy-six controls were scanned once without medication to provide normative data. Seed-based correlation analyses were used to measure changes in functional connectivity, with the right inferior frontal gyrus (IFG) a critical region for executive function. Patients on placebo had reduced connectivity relative to controls from right IFG to dorsal anterior cingulate cortex and to left IFG and dorsolateral prefrontal cortex. Atomoxetine increased connectivity from the right IFG to the dorsal anterior cingulate. In addition, the atomoxetine-induced change in connectivity from right IFG to dorsolateral prefrontal cortex was proportional to the change in verbal fluency, a simple index of executive function. The results support the hypothesis that atomoxetine may restore prefrontal networks related to executive functions. We suggest that task-free imaging can support translational pharmacological studies of new drug therapies and provide evidence for engagement of the relevant neurocognitive systems.

Cortical mechanisms of mirror therapy after stroke.

BACKGROUND AND OBJECTIVE: Mirror therapy is a new form of stroke rehabilitation that uses the mirror reflection of the unaffected hand in place of the affected hand to augment movement training. The mechanism of mirror therapy is not known but is thought to involve changes in cerebral organization. We used magnetoencephalography (MEG) to measure changes in cortical activity during mirror training after stroke. In particular, we examined movement-related changes in the power of cortical oscillations in the beta (15-30 Hz) frequency range, known to be involved in movement. METHODS: Ten stroke patients with upper limb paresis and 13 healthy controls were recorded using MEG while performing bimanual hand movements in 2 different conditions. In one, subjects looked directly at their affected hand (or dominant hand in controls), and in the other, they looked at a mirror reflection of their unaffected hand in place of their affected hand. The movement-related beta desynchronization was calculated in both primary motor cortices. RESULTS: Movement-related beta desynchronization was symmetrical during bilateral movement and unaltered by the mirror condition in controls. In the patients, movement-related beta desynchronization was generally smaller than in controls, but greater in contralesional compared to ipsilesional motor cortex. This initial asymmetry in movement-related beta desynchronization between hemispheres was made more symmetrical by the presence of the mirror. CONCLUSIONS: Mirror therapy could potentially aid stroke rehabilitation by normalizing an asymmetrical pattern of movement-related beta desynchronization in primary motor cortices during bilateral movement.