Automated MRI perfusion-diffusion mismatch estimation may be significantly different in individual patients when using different software packages.

OBJECTIVE: To compare two established software applications in terms of apparent diffusion coefficient (ADC) lesion volumes, volume of critically hypoperfused brain tissue, and calculated volumes of perfusion-diffusion mismatch in brain MRI of patients with acute ischemic stroke. METHODS: Brain MRI examinations of 81 patients with acute stroke due to large vessel occlusion of the anterior circulation were analyzed. The volume of hypoperfused brain tissue, ADC volume, and the volume of perfusion-diffusion mismatch were calculated automatically with two different software packages. The calculated parameters were compared quantitatively using formal statistics. RESULTS: Significant difference was found for the volume of hypoperfused tissue (median 91.0 ml vs. 102.2 ml; p < 0.05) and the ADC volume (median 30.0 ml vs. 23.9 ml; p < 0.05) between different software packages. The volume of the perfusion-diffusion mismatch differed significantly (median 47.0 ml vs. 67.2 ml; p < 0.05). Evaluation of the results on a single-subject basis revealed a mean absolute difference of 20.5 ml for hypoperfused tissue, 10.8 ml for ADC volumes, and 27.6 ml for mismatch volumes, respectively. Application of the DEFUSE 3 threshold of 70 ml infarction core would have resulted in dissenting treatment decisions in 6/81 (7.4%) patients. CONCLUSION: Volume segmentation in different software products may lead to significantly different results in the individual patient and may thus seriously influence the decision for or against mechanical thrombectomy. KEY POINTS: • Automated calculation of MRI perfusion-diffusion mismatch helps clinicians to apply inclusion and exclusion criteria derived from randomized trials. • Infarct volume segmentation plays a crucial role and lead to significantly different result for different computer programs. • Perfusion-diffusion mismatch estimation from different computer programs may influence the decision for or against mechanical thrombectomy.

Interference and problem size effect in multiplication fact solving: Individual differences in brain activations and arithmetic performance.

In the development of math ability, a large variability of performance in solving simple arithmetic problems is observed and has not found a compelling explanation yet. One robust effect in simple multiplication facts is the problem size effect, indicating better performance for small problems compared to large ones. Recently, behavioral studies brought to light another effect in multiplication facts, the interference effect. That is, high interfering problems (receiving more proactive interference from previously learned problems) are more difficult to retrieve than low interfering problems (in terms of physical feature overlap, namely the digits, De Visscher and Noël, 2014). At the behavioral level, the sensitivity to the interference effect is shown to explain individual differences in the performance of solving multiplications in children as well as in adults. The aim of the present study was to investigate the individual differences in multiplication ability in relation to the neural interference effect and the neural problem size effect. To that end, we used a paradigm developed by De Visscher, Berens, et al. (2015) that contrasts the interference effect and the problem size effect in a multiplication verification task, during functional magnetic resonance imaging (fMRI) acquisition. Forty-two healthy adults, who showed high variability in an arithmetic fluency test, participated in our fMRI study. In order to control for the general reasoning level, the IQ was taken into account in the individual differences analyses. Our findings revealed a neural interference effect linked to individual differences in multiplication in the left inferior frontal gyrus, while controlling for the IQ. This interference effect in the left inferior frontal gyrus showed a negative relation with individual differences in arithmetic fluency, indicating a higher interference effect for low performers compared to high performers. This region is suggested in the literature to be involved in resolution of proactive interference. Besides, no correlation between the neural problem size effect and multiplication performance was found. This study supports the idea that the interference due to similarities/overlap of physical traits (the digits) is crucial in memorizing arithmetic facts and in determining individual differences in arithmetic.

The left intraparietal sulcus adapts to symbolic number in both the visual and auditory modalities: Evidence from fMRI.

A growing body of evidence from functional Magnetic Resonance Imaging adaptation (fMRIa) has implicated the left intraparietal sulcus (IPS) as a crucial brain region representing the semantic of number symbols. However, it is currently unknown to what extent the left IPS brain activity can be generalized across modalities (e.g., Arabic digits and spoken number words) and how robust and reproducible numerical adaptation effects are. In two separate fMRIa experiments we habituated the brain response of 20 native English-speaking (Experiment 1) and 34 native German-speaking (Experiment 2) adults to Arabic digits or spoken number words. Consistent with previous findings, experiment 1 revealed numerical ratio dependent adaptation to Arabic numerals in the left IPS using both conventional and cortex-based alignment techniques. Experiment 2 revealed numerical ratio dependent signal recovery in the left IPS following adaptation to both Arabic numerals and spoken number words using both conventional and cortex-based alignment techniques. Together, these findings suggest that the left IPS is involved in symbolic number processing across modalities.

In vivo cardiovascular magnetic resonance of 2D vessel wall diffusion anisotropy in carotid arteries.

BACKGROUND: Diffusion weighted (DW) cardiovascular magnetic resonance (CMR) has shown great potential to discriminate between healthy and diseased vessel tissue by evaluating the apparent diffusion coefficient (ADC) along the arterial axis. Recently, ex vivo studies on porcine arteries utilizing diffusion tensor imaging (DTI) revealed a circumferential fiber orientation rather than an organization in axial direction, suggesting dominant diffusion perpendicular to the slice direction. In the present study, we propose a method to access tangential and radial diffusion of carotids in vivo by utilizing a pulse sequence that enables high resolution DW imaging in combination with a two-dimensional (2D) diffusion gradient direction sampling scheme perpendicular to the longitudinal axis of the artery. METHODS: High resolution DTI of 12 healthy male volunteers (age: 25-60 years) was performed on one selected axial slice using a read-out segmented EPI (rs-EPI) sequence on a 3T MR scanner. RESULTS: It was found consistently for all 12 volunteers, that the tangential component as the principle direction of diffusion. Mean vessel wall fractional anisotropy (FA) values ranged from 0.7 for the youngest to 0.56 for the oldest participant. Linear regression analysis between the FA values and volunteers age revealed a highly significant (P < 0.01) linear relationship with an adjusted R2 of 0.52. In addition, a linear trend (P < 0.1) could be observed between radial diffusivity (RD) and age. CONCLUSION: These results point to FA being a sensitive parameter able to capture changes in the vascular architecture with age. In detail, the data demonstrate a decrease in FA with advancing age indicating possible alterations of tissue microstructural integrity. Moreover, analyzing 2D diffusion tensor directions is sufficient and applicable in a clinical setup concerning the overall scan time.

Fast quantitative susceptibility mapping using 3D EPI and total generalized variation.

Quantitative susceptibility mapping (QSM) allows new insights into tissue composition and organization by assessing its magnetic property. Previous QSM studies have already demonstrated that magnetic susceptibility is highly sensitive to myelin density and fiber orientation as well as to para- and diamagnetic trace elements. Image resolution in QSM with current approaches is limited by the long acquisition time of 3D scans and the need for high signal to noise ratio (SNR) to solve the dipole inversion problem. We here propose a new total-generalized-variation (TGV) based method for QSM reconstruction, which incorporates individual steps of phase unwrapping, background field removal and dipole inversion in a single iteration, thus yielding a robust solution to the reconstruction problem. This approach has beneficial characteristics for low SNR data, allowing for phase data to be rapidly acquired with a 3D echo planar imaging (EPI) sequence. The proposed method was evaluated with a numerical phantom and in vivo at 3 and 7 T. Compared to total variation (TV), TGV-QSM enforced higher order smoothness which yielded solutions closer to the ground truth and prevented stair-casing artifacts. The acquisition time for images with 1mm isotropic resolution and whole brain coverage was 10s on a clinical 3 Tesla scanner. In conclusion, 3D EPI acquisition combined with single-step TGV reconstruction yields reliable QSM images of the entire brain with 1mm isotropic resolution in seconds. The short acquisition time combined with the robust reconstruction may enable new QSM applications in less compliant populations, clinical susceptibility tensor imaging, and functional resting state examinations.

To create or to recall? Neural mechanisms underlying the generation of creative new ideas.

This fMRI study investigated brain activation during creative idea generation using a novel approach allowing spontaneous self-paced generation and expression of ideas. Specifically, we addressed the fundamental question of what brain processes are relevant for the generation of genuinely new creative ideas, in contrast to the mere recollection of old ideas from memory. In general, creative idea generation (i.e., divergent thinking) was associated with extended activations in the left prefrontal cortex and the right medial temporal lobe, and with deactivation of the right temporoparietal junction. The generation of new ideas, as opposed to the retrieval of old ideas, was associated with stronger activation in the left inferior parietal cortex which is known to be involved in mental simulation, imagining, and future thought. Moreover, brain activation in the orbital part of the inferior frontal gyrus was found to increase as a function of the creativity (i.e., originality and appropriateness) of ideas pointing to the role of executive processes for overcoming dominant but uncreative responses. We conclude that the process of idea generation can be generally understood as a state of focused internally-directed attention involving controlled semantic retrieval. Moreover, left inferior parietal cortex and left prefrontal regions may subserve the flexible integration of previous knowledge for the construction of new and creative ideas.

Creativity and schizotypy from the neuroscience perspective.

Behavioral research has revealed that some cognitive features may be similar between creative and psychotic/schizophrenic-like thoughts. In this study, we addressed the potential link between creativity and schizotypy at the level of the brain by investigating functional patterns of brain activity (using functional magnetic resonance imaging) during creative cognition in preselected groups with low versus high psychometrically determined schizotypy. Our findings revealed an association between the originality component of creativity and reduced deactivation of right parietal brain regions and the precuneus during creative cognition, congruent with the idea that more-creative people may include many more events/stimuli in their mental processes than do less-creative people. Similarly, the high-schizotypy group showed weaker deactivation of the right precuneus during creative cognition. The fact that originality and schizotypy show similar functional brain activity patterns during creative ideation (i.e., reduced deactivation of the right precuneus) strongly supports the contention that similar mental processes may be implicated in creativity and in psychosis proneness.

Fractal-Based Analysis of Arteriovenous Malformations (AVMs).

Arteriovenous malformations (AVMs) are cerebrovascular lesions consisting of a pathologic tangle of the vessels characterized by a core termed the nidus, which is the "nest" where the fistulous connections occur. AVMs can cause headache, stroke, and/or seizures. Their treatment can be challenging requiring surgery, endovascular embolization, and/or radiosurgery as well. AVMs' morphology varies greatly among patients, and there is still a lack of standardization of angioarchitectural parameters, which can be used as morphometric parameters as well as potential clinical biomarkers (e.g., related to prognosis).In search of new diagnostic and prognostic neuroimaging biomarkers of AVMs, computational fractal-based models have been proposed for describing and quantifying the angioarchitecture of the nidus. In fact, the fractal dimension (FD) can be used to quantify AVMs' branching pattern. Higher FD values are related to AVMs characterized by an increased number and tortuosity of the intranidal vessels or to an increasing angioarchitectural complexity as a whole. Moreover, FD has been investigated in relation to the outcome after Gamma Knife radiosurgery, and an inverse relationship between FD and AVM obliteration was found.Taken altogether, FD is able to quantify in a single and objective value what neuroradiologists describe in qualitative and/or semiquantitative way, thus confirming FD as a reliable morphometric neuroimaging biomarker of AVMs and as a potential surrogate imaging biomarker. Moreover, computational fractal-based techniques are under investigation for the automatic segmentation and extraction of the edges of the nidus in neuroimaging, which can be relevant for surgery and/or radiosurgery planning.

Wall Enhancement of Coiled Intracranial Aneurysms Is Associated with Aneurysm Recanalization: A Cross-Sectional Study.

BACKGROUND AND PURPOSE: Wall enhancement of untreated intracranial aneurysms on MR imaging is thought to predict aneurysm instability. Wall enhancement or enhancement of the aneurysm cavity in coiled intracranial aneurysms is discussed controversially in the literature regarding potential healing mechanisms or adverse inflammatory reactions. Our aim was to compare the occurrence of aneurysm wall enhancement and cavity enhancement between completely occluded intracranial aneurysms and recanalized aneurysms after initially complete coil embolization. MATERIALS AND METHODS: In this single-center cross-sectional study, we evaluated intracranial aneurysms after successful coil embolization for aneurysm recanalization, wall enhancement, and cavity enhancement with 3T MR imaging. We then compared the incidence of wall enhancement and cavity enhancement of completely occluded aneurysms with aneurysms with recanalization using the χ2 test and performed a multivariate linear regression analysis with recanalization size as an independent variable. RESULTS: We evaluated 59 patients (mean age, 54.7 [SD, 12.4] years; 48 women) with 60 intracranial aneurysms and found a significantly higher incidence of wall enhancement in coiled aneurysms with recanalization (n=38) compared with completely occluded aneurysms (n = 22, P = .036). In addition, there was a significantly higher incidence of wall enhancement in aneurysms with recanalization of >3 mm (P = .003). In a multivariate linear regression analysis, wall enhancement (P = .010) and an increase of overall aneurysm size after embolization (P < .001) were significant predictors of recanalization size (corrected R 2= 0.430, CI 95%). CONCLUSIONS: The incidence of aneurysm wall enhancement is increased in coiled intracranial aneurysms with recanalization and is associated with recanalization size.

White matter integrity and functional connectivity of the default mode network in acute stroke are associated with cognitive outcome three months post-stroke.

BACKGROUND: Knowledge about factors that are associated with post-stroke cognitive outcome is important to identify patients with high risk for impairment. We therefore investigated the associations of white matter integrity and functional connectivity (FC) within the brain's default-mode network (DMN) in acute stroke patients with cognitive outcome three months post-stroke. METHODS: Patients aged between 18 and 85 years with an acute symptomatic MRI-proven unilateral ischemic middle cerebral artery infarction, who had received reperfusion therapy, were invited to participate in this longitudinal study. All patients underwent brain MRI within 24-72 h after symptom onset, and participated in a neuropsychological assessment three months post-stroke. We performed hierarchical regression analyses to explore the incremental value of baseline white matter integrity and FC beyond demographic, clinical, and macrostructural information for cognitive outcome. RESULTS: The study cohort comprised 34 patients (mean age: 64 ± 12 years, 35% female). The initial median National Institutes of Health Stroke Scale (NIHSS) score was 10, and significantly improved three months post-stroke to a median NIHSS = 1 (p < .001). Nonetheless, 50% of patients showed cognitive impairment three months post-stroke. FC of the non-lesioned anterior cingulate cortex of the affected hemisphere explained 15% of incremental variance for processing speed (p = .007), and fractional anisotropy of the non-lesioned cingulum of the affected hemisphere explained 13% of incremental variance for cognitive flexibility (p = .033). CONCLUSIONS: White matter integrity and functional MRI markers of the DMN in acute stroke explain incremental variance for post-stroke cognitive outcome beyond demographic, clinical, and macrostructural information.

The function of the left angular gyrus in mental arithmetic: evidence from the associative confusion effect.

While the left angular gyrus (lAG) has been repeatedly implicated in mental arithmetic, its precise functional role has not been established. On the one hand, it has been speculated that the lAG is involved in task-specific processes. On the other hand, the observation of relative deactivation during arithmetic has led to the contention that differential lAG activation reflects task-unrelated difficulty effects associated with the default mode network (DMN). Using functional magnetic resonance imaging, we investigated the neural correlates of the associative confusion effect that allowed us to dissociate effects of task difficulty and task-related arithmetic processes on lAG activation. The associative confusion effect is characterized by poorer performance while verifying addition and multiplication equations whose solutions are associated with the other operation (confusion equations: e.g., "9 × 6 = 15") compared with solutions unrelated to both operations (non-confusion equations: e.g., "9 × 6 = 52"). Comparing these two conditions revealed higher activation of the anterior lAG (areas PGa, PFm, and PF) and the left dorsolateral prefrontal cortex for the confusion problems. This effect displayed only slight anatomical overlap with the well-established reverse problem-size effect (small minus large problems) and task-related deactivation in the parietal cortex. The finding of greater lAG activity (less deactivation) in the more difficult task condition is inconsistent with the hypothesis that lAG activation during mental arithmetic reflects task difficulty related modulations of the DMN. Instead, the present findings provide further support for the symbol-referent mapping hypothesis, suggesting that the lAG mediates the automatic mapping of arithmetic problems onto solutions stored in memory.

Individual differences in solving arithmetic word problems.

BACKGROUND: With the present functional magnetic resonance imaging (fMRI) study at 3 T, we investigated the neural correlates of visualization and verbalization during arithmetic word problem solving. In the domain of arithmetic, visualization might mean to visualize numbers and (intermediate) results while calculating, and verbalization might mean that numbers and (intermediate) results are verbally repeated during calculation. If the brain areas involved in number processing are domain-specific as assumed, that is, that the left angular gyrus (AG) shows an affinity to the verbal domain, and that the left and right intraparietal sulcus (IPS) shows an affinity to the visual domain, the activation of these areas should show a dependency on an individual's cognitive style. METHODS: 36 healthy young adults participated in the fMRI study. The participants habitual use of visualization and verbalization during solving arithmetic word problems was assessed with a short self-report assessment. During the fMRI measurement, arithmetic word problems that had to be solved by the participants were presented in an event-related design. RESULTS: We found that visualizers showed greater brain activation in brain areas involved in visual processing, and that verbalizers showed greater brain activation within the left angular gyrus. CONCLUSIONS: Our results indicate that cognitive styles or preferences play an important role in understanding brain activation. Our results confirm, that strong visualizers use mental imagery more strongly than weak visualizers during calculation. Moreover, our results suggest that the left AG shows a specific affinity to the verbal domain and subserves number processing in a modality-specific way.

Information extraction from German radiological reports for general clinical text and language understanding.

Recent advances in deep learning and natural language processing (NLP) have opened many new opportunities for automatic text understanding and text processing in the medical field. This is of great benefit as many clinical downstream tasks rely on information from unstructured clinical documents. However, for low-resource languages like German, the use of modern text processing applications that require a large amount of training data proves to be difficult, as only few data sets are available mainly due to legal restrictions. In this study, we present an information extraction framework that was initially pre-trained on real-world computed tomographic (CT) reports of head examinations, followed by domain adaptive fine-tuning on reports from different imaging examinations. We show that in the pre-training phase, the semantic and contextual meaning of one clinical reporting domain can be captured and effectively transferred to foreign clinical imaging examinations. Moreover, we introduce an active learning approach with an intrinsic strategic sampling method to generate highly informative training data with low human annotation cost. We see that the model performance can be significantly improved by an appropriate selection of the data to be annotated, without the need to train the model on a specific downstream task. With a general annotation scheme that can be used not only in the radiology field but also in a broader clinical setting, we contribute to a more consistent labeling and annotation process that also facilitates the verification and evaluation of language models in the German clinical setting.

Workflow assessment of an augmented reality application for planning of perforator flaps in plastic reconstructive surgery: Game or game changer?

Objective: In contrast to the rising amount of financial investments for research and development in medical technology worldwide is the lack of usability and clinical readiness of the produced systems. We evaluated an augmented reality (AR) setup under development for preoperative perforator vessel mapping for elective autologous breast reconstruction. Methods: In this grant-supported research pilot, we used magnetic resonance angiography data (MR-A) of the trunk to superimpose the scans on the corresponding patients with hands-free AR goggles to identify regions-of-interest for surgical planning. Perforator location was assessed using MR-A imaging (MR-A projection) and Doppler ultrasound data (3D distance) and confirmed intraoperatively in all cases. We evaluated usability (System Usability Scale, SUS), data transfer load and documented personnel hours for software development, correlation of image data, as well as processing duration to clinical readiness (time from MR-A to AR projections per scan). Results: All perforator locations were confirmed intraoperatively, and we found a strong correlation between MR-A projection and 3D distance measurements (Spearman r = 0.894). The overall usability (SUS) was 67 ± 10 (=moderate to good). The presented setup for AR projections took 173 min to clinical readiness (=availability on AR device per patient). Conclusion: In this pilot, we calculated development investments based on project-approved grant-funded personnel hours with a moderate to good usability outcome resulting from some limitations: assessment was based on one-time testing with no previous training, a time lag of AR visualizations on the body and difficulties in spatial AR orientation. The use of AR systems can provide new opportunities for future surgical planning, but has more potential for educational (e.g., patient information) or training purposes of medical under- and postgraduates (spatial recognition of imaging data associated with anatomical structures and operative planning). We expect future usability improvements with refined user interfaces, faster AR hardware and artificial intelligence-enhanced visualization techniques.

Quantitative susceptibility mapping (QSM) as a means to measure brain iron? A post mortem validation study.

Quantitative susceptibility mapping (QSM) is a novel technique which allows determining the bulk magnetic susceptibility distribution of tissue in vivo from gradient echo magnetic resonance phase images. It is commonly assumed that paramagnetic iron is the predominant source of susceptibility variations in gray matter as many studies have reported a reasonable correlation of magnetic susceptibility with brain iron concentrations in vivo. Instead of performing direct comparisons, however, all these studies used the putative iron concentrations reported in the hallmark study by Hallgren and Sourander (1958) for their analysis. Consequently, the extent to which QSM can serve to reliably assess brain iron levels is not yet fully clear. To provide such information we investigated the relation between bulk tissue magnetic susceptibility and brain iron concentration in unfixed (in situ) post mortem brains of 13 subjects using MRI and inductively coupled plasma mass spectrometry. A strong linear correlation between chemically determined iron concentration and bulk magnetic susceptibility was found in gray matter structures (r=0.84, p<0.001), whereas the correlation coefficient was much lower in white matter (r=0.27, p<0.001). The slope of the overall linear correlation was consistent with theoretical considerations of the magnetism of ferritin supporting that most of the iron in the brain is bound to ferritin proteins. In conclusion, iron is the dominant source of magnetic susceptibility in deep gray matter and can be assessed with QSM. In white matter regions the estimation of iron concentrations by QSM is less accurate and more complex because the counteracting contribution from diamagnetic myelinated neuronal fibers confounds the interpretation.

Stimulating creativity via the exposure to other people's ideas.

As it is the case in brainstorming, each single idea a person generates to a specific problem may stimulate new ideas or solutions in others. In this fMRI study, we investigate the effects of cognitive stimulation via the exposure to other people's ideas on the originality of generated ideas. Participants are requested to generate alternative uses of conventional everyday objects subsequent to a short cognitive stimulation intervention in which they are exposed to other ideas, which were either common or highly original. In a control condition, meaningless pseudowords are shown. Results suggest that cognitive stimulation via common or moderately creative ideas was effective in improving creativity. At the neurophysiological level, temporo-parietal brain regions (primarily right-hemispheric) turned out to be particularly sensitive to cognitive stimulation, possibly indicating that cognitive stimulation via relevant memory cues results in a state of heightened focused attention to memory that facilitates efficient retrieval and recombination of existing knowledge.

The Influence of verbalization on the pattern of cortical activation during mental arithmetic.

BACKGROUND: The aim of the present functional magnetic resonance imaging (fMRI) study at 3 T was to investigate the influence of the verbal-visual cognitive style on cerebral activation patterns during mental arithmetic. In the domain of arithmetic, a visual style might for example mean to visualize numbers and (intermediate) results, and a verbal style might mean, that numbers and (intermediate) results are verbally repeated. In this study, we investigated, first, whether verbalizers show activations in areas for language processing, and whether visualizers show activations in areas for visual processing during mental arithmetic. Some researchers have proposed that the left and right intraparietal sulcus (IPS), and the left angular gyrus (AG), two areas involved in number processing, show some domain or modality specificity. That is, verbal for the left AG, and visual for the left and right IPS. We investigated, second, whether the activation in these areas implied in number processing depended on an individual's cognitive style. METHODS: 42 young healthy adults participated in the fMRI study. The study comprised two functional sessions. In the first session, subtraction and multiplication problems were presented in an event-related design, and in the second functional session, multiplications were presented in two formats, as Arabic numerals and as written number words, in an event-related design. The individual's habitual use of visualization and verbalization during mental arithmetic was assessed by a short self-report assessment. RESULTS: We observed in both functional sessions that the use of verbalization predicts activation in brain areas associated with language (supramarginal gyrus) and auditory processing (Heschl's gyrus, Rolandic operculum). However, we found no modulation of activation in the left AG as a function of verbalization. CONCLUSIONS: Our results confirm that strong verbalizers use mental speech as a form of mental imagination more strongly than weak verbalizers. Moreover, our results suggest that the left AG has no specific affinity to the verbal domain and subserves number processing in a modality-general way.

Reproducibility of artificial intelligence models in computed tomography of the head: a quantitative analysis.

When developing artificial intelligence (AI) software for applications in radiology, the underlying research must be transferable to other real-world problems. To verify to what degree this is true, we reviewed research on AI algorithms for computed tomography of the head. A systematic review was conducted according to the preferred reporting items for systematic reviews and meta-analyses. We identified 83 articles and analyzed them in terms of transparency of data and code, pre-processing, type of algorithm, architecture, hyperparameter, performance measure, and balancing of dataset in relation to epidemiology. We also classified all articles by their main functionality (classification, detection, segmentation, prediction, triage, image reconstruction, image registration, fusion of imaging modalities). We found that only a minority of authors provided open source code (10.15%, n 0 7), making the replication of results difficult. Convolutional neural networks were predominantly used (32.61%, n = 15), whereas hyperparameters were less frequently reported (32.61%, n = 15). Data sets were mostly from single center sources (84.05%, n = 58), increasing the susceptibility of the models to bias, which increases the error rate of the models. The prevalence of brain lesions in the training (0.49 ± 0.30) and testing (0.45 ± 0.29) datasets differed from real-world epidemiology (0.21 ± 0.28), which may overestimate performances. This review highlights the need for open source code, external validation, and consideration of disease prevalence.

Effects of actual and imagined music-cued gait training on motor functioning and brain activity in people with multiple sclerosis: protocol of a randomised parallel multicentre trial.

INTRODUCTION: Motor imagery (MI) refers to the mental rehearsal of a physical action without muscular activity. Our previous studies showed that MI combined with rhythmic-auditory cues improved walking, fatigue and quality of life (QoL) in people with multiple sclerosis (pwMS). Largest improvements were seen after music and verbally cued MI. It is unclear whether actual cued gait training achieves similar effects on walking as cued MI in pwMS. Furthermore, in pwMS it is unknown whether any of these interventions leads to changes in brain activation. The purpose of this study is therefore to compare the effects of imagined and actual cued gait training and a combination thereof on walking, brain activation patterns, fatigue, cognitive and emotional functioning in pwMS. METHODS AND ANALYSIS: A prospective double-blind randomised parallel multicentre trial will be conducted in 132 pwMS with mild to moderate disability. Randomised into three groups, participants will receive music, metronome and verbal cueing, plus MI of walking (1), MI combined with actual gait training (2) or actual gait training (3) for 30 min, 4× per week for 4 weeks. Supported by weekly phone calls, participants will practise at home, guided by recorded instructions. Primary endpoints will be walking speed (Timed 25-Foot Walk) and distance (2 min Walk Test). Secondary endpoints will be brain activation patterns, fatigue, QoL, MI ability, anxiety, depression, cognitive functioning, music-induced motivation-to-move, pleasure, arousal and self-efficacy. Data will be collected at baseline, postintervention and 3-month follow-up. MRI reference values will be generated using 15 matched healthy controls. ETHICS AND DISSEMINATION: This study follows the Standard Protocol Items: Recommendations for Interventional Trials-PRO Extension. Ethical approval was received from the Ethics Committees of the Medical Universities of Innsbruck (1347/2020) and Graz (33-056 ex 20/21), Austria. Results will be disseminated via national and international conferences and published in peer-reviewed journals. TRIAL REGISTRATION NUMBER: DRKS00023978.