Random forest-based prediction of stroke outcome.

We research into the clinical, biochemical and neuroimaging factors associated with the outcome of stroke patients to generate a predictive model using machine learning techniques for prediction of mortality and morbidity 3-months after admission. The dataset consisted of patients with ischemic stroke (IS) and non-traumatic intracerebral hemorrhage (ICH) admitted to Stroke Unit of a European Tertiary Hospital prospectively registered. We identified the main variables for machine learning Random Forest (RF), generating a predictive model that can estimate patient mortality/morbidity according to the following groups: (1) IS + ICH, (2) IS, and (3) ICH. A total of 6022 patients were included: 4922 (mean age 71.9 ± 13.8 years) with IS and 1100 (mean age 73.3 ± 13.1 years) with ICH. NIHSS at 24, 48 h and axillary temperature at admission were the most important variables to consider for evolution of patients at 3-months. IS + ICH group was the most stable for mortality prediction [0.904 ± 0.025 of area under the receiver operating characteristics curve (AUC)]. IS group presented similar results, although variability between experiments was slightly higher (0.909 ± 0.032 of AUC). ICH group was the one in which RF had more problems to make adequate predictions (0.9837 vs. 0.7104 of AUC). There were no major differences between IS and IS + ICH groups according to morbidity prediction (0.738 and 0.755 of AUC) but, after checking normality with a Shapiro Wilk test with the null hypothesis that the data follow a normal distribution, it was rejected with W = 0.93546 (p-value < 2.2e-16). Conditions required for a parametric test do not hold, and we performed a paired Wilcoxon Test assuming the null hypothesis that all the groups have the same performance. The null hypothesis was rejected with a value < 2.2e-16, so there are statistical differences between IS and ICH groups. In conclusion, machine learning algorithms RF can be effectively used in stroke patients for long-term outcome prediction of mortality and morbidity.

Predicting conversion to multiple sclerosis by assessing cognitive impairment in radiologically isolated syndrome.

Up to a third of patients with radiologically isolated syndrome (RIS) exhibit lower-than-expected cognitive performances in neuropsychological evaluations, but the relationship between cognitive impairment (CI) and quantitative magnetic resonance (MRI) measures has not been stablished. Furthermore, the prognostic role of CI in RIS for conversion to MS is currently unknown. We assessed 17 patients with RIS and 17 matched healthy controls (HC) with a neurophychological battery and a 3T MRI. Six patients (35,3%) fulfilled our criterion for CI (scores 2 SDs below the mean of HC in at least two cognitive tests) (ci-RIS). The ci-RIS subgroup showed lower values of normalized brain and gray matter volumes when compared to HC. After a median follow-up time of 4.5 years, the ci-RIS subgroup presented a higher conversion rate to MS, suggesting that CI might be an independent risk factor for conversion to MS.

A data mining approach for classification of orthostatic and essential tremor based on MRI-derived brain volume and cortical thickness.

OBJECTIVE: Orthostatic tremor (OT) is an extremely rare, misdiagnosed, and underdiagnosed disorder affecting adults in midlife. There is debate as to whether it is a different condition or a variant of essential tremor (ET), or even, if both conditions coexist. Our objective was to use data mining classification methods, using magnetic resonance imaging (MRI)-derived brain volume and cortical thickness data, to identify morphometric measures that help to discriminate OT patients from those with ET. METHODS: MRI-derived brain volume and cortical thickness were obtained from 14 OT patients and 15 age-, sex-, and education-matched ET patients. Feature selection and machine learning methods were subsequently applied. RESULTS: Four MRI features alone distinguished the two, OT from ET, with 100% diagnostic accuracy. More specifically, left thalamus proper volume (normalized by the total intracranial volume), right superior parietal volume, right superior parietal thickness, and right inferior parietal roughness (i.e., the standard deviation of cortical thickness) were shown to play a key role in OT and ET characterization. Finally, the left caudal anterior cingulate thickness and the left caudal middle frontal roughness allowed us to separate with 100% diagnostic accuracy subgroups of OT patients (primary and those with mild parkinsonian signs). CONCLUSIONS: A data mining approach applied to MRI-derived brain volume and cortical thickness data may differentiate between these two types of tremor with an accuracy of 100%. Our results suggest that OT and ET are distinct conditions.

Diffusion tensor imaging in orthostatic tremor: a tract-based spatial statistics study.

OBJECTIVE: The pathogenesis of orthostatic tremor (OT) is unknown. We investigated OT-related white matter changes and their correlations with scores from a neuropsychological testing battery. METHODS: Diffusion tensor imaging measures were compared between 14 OT patients and 14 age- and education-matched healthy controls, using whole-brain tract-based spatial statistics analysis. Correlations between altered diffusion metrics and cognitive performance in OT group were assessed. RESULTS: In all cognitive domains (attention, executive function, visuospatial ability, verbal memory, visual memory, and language), OT patients' cognitive performance was significantly worse than that of healthy controls. OT patients demonstrated altered diffusivity metrics not only in the posterior lobe of the cerebellum (left cerebellar lobule VI) and in its efferent cerebellar fibers (left superior cerebellar peduncle), but also in medial lemniscus bilaterally (pontine tegmentum), anterior limb of the internal capsule bilaterally, right posterior limb of the internal capsule, left anterior corona radiata, right insula, and the splenium of corpus callosum. No relationship was found between diffusion measures and disease duration in OT patients. Diffusion white matter changes, mainly those located in right anterior limb of the internal capsule, were correlated with poor performance on tests of executive function, visuospatial ability, verbal memory, and visual memory in OT patients. INTERPRETATION: White matter changes were preferentially located in the cerebellum, its efferent pathways, as well as in the pontine tegmentum and key components of the frontal-thalamic-cerebellar circuit. Further work needs to be done to understand the evolution of these white matter changes and their functional consequences.

Author Correction: White matter microstructural changes are related to cognitive dysfunction in essential tremor.

A correction to this article has been published and is linked from the HTML and PDF versions of this paper. The error has not been fixed in the paper.

White matter microstructural changes are related to cognitive dysfunction in essential tremor.

Diffusion tensor imaging (DTI) studies have detected white matter microstructural changes in essential tremor (ET). However, it is still unclear whether these changes are related to cognitive deficits, which have been described in ET patients. DTI-derived fractional anisotropy, mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity measures were compared between 23 ET patients and 23 age-, gender-, and education-matched healthy individuals, using whole-brain tract-based spatial statistics. Correlations of white matter changes with scores obtained from a detailed neuropsychological assessment were subsequently examined. ET patients demonstrated increases in MD in the bilateral posterior corona radiata, bilateral superior longitudinal fasciculus, bilateral fornix (cres)/stria terminalis, genu and splenium of the corpus callosum, bilateral anterior and posterior limbs of internal capsule, bilateral retrolenticular region part of internal capsule, and left posterior thalamic radiation. Except for the genu of the corpus callosum, an increase in AD values was also found in these same tracts. Furthermore, increased MD and AD values in different white matter areas was negatively correlated with performance on language and verbal memory and positively with visuospatial ability. These correlations suggest that white matter changes might be involved in the pathogenesis of cognitive deficits in ET.

In vivo neurometabolic profiling in orthostatic tremor.

The pathogenesis of orthostatic tremor (OT) remains unclear, although some evidence points to dysfunction in the brainstem or cerebellum. We used single voxel proton magnetic resonance spectroscopy (1H-MRS) (3 T) to investigate whether neurochemical changes underlie abnormal cerebellar or cortical function in OT. Fourteen OT patients and 14 healthy controls underwent 1H-MRS studies with voxels placed in midparietal gray matter and cerebellum (vermis and central white matter). Spectral analysis was analyzed using the software package LCModel (version 6.3). The absolute metabolite concentrations and ratios of total N-acetylaspartate + N-acetylaspartyl glutamate (NAA), choline-containing compounds, myoinositol, and glutamate + glutamine to creatine were calculated. In midparietal gray matter spectra, we found a significant decrease in the absolute concentration of NAA in OT patients versus healthy controls (7.76 ±â€Š0.25 vs 8.11 ±â€Š0.45, P = 0.017). A similar decrease in NAA was seen in the cerebellar vermis (7.33 ±â€Š0.61 vs 8.55 ±â€Š1.54, P = 0.014) and cerebellar white matter (8.54 ±â€Š0.79 vs 9.95 ±â€Š1.57, P = 0.010). No differences in the other metabolites or their ratios were observed. Reductions in both cerebral cortical and cerebellar NAA suggest that there is neuronal damage or loss in OT, raising the intriguing question as to whether OT is a neurodegenerative disease. Along with clinical history and electrophysio0logical examination, 1H-MRS could serve as a useful diagnostic aid for OT.

Normal-appearing brain tissue analysis in radiologically isolated syndrome using 3 T MRI.

To date, it remains largely unknown whether there is in radiologically isolated syndrome (RIS) brain damage beyond visible T2 white matter lesions. We used single- voxel proton magnetic resonance spectroscopy and diffusion tensor imaging (3 T MRI) to analyze normal-appearing brain tissue regions in 18 RIS patients and 18 matched healthy controls. T2-hyperintense lesion volumes and structural brain volumes were also measured. The absolute metabolite concentrations and ratios of total N-acetylaspartate+N-acetylaspartyl glutamate (NAA), choline-containing compounds, myoinositol, and glutamine-glutamate complex to creatine were calculated. Spectral analysis was performed by LCModel. Voxelwise morphometry analysis was performed to localize regions of brain tissue showing significant changes of fractional anisotropy or mean diffusivity. Compared with healthy controls, RIS patients did not show any significant differences in either the absolute concentration of NAA or NAA/Cr ratio in mid-parietal gray matter. A trend toward lower NAA concentrations (-3.35%) was observed among RIS patients with high risk for conversion to multiple sclerosis. No differences in the other metabolites or their ratios were observed. RIS patients showed lower fractional anisotropy only in clusters overlapping lesional areas, namely in the cingulate gyrus bilaterally and the frontal lobe subgyral bilaterally (P < 0.001). Normalized brain and cortical volumes were significantly lower in RIS patients than in controls (P = 0.01 and P = 0.03, respectively). Our results suggest that in RIS, global brain and cortical atrophy are not primarily driven by significant occult microstructural normal appearing brain damage. Longitudinal MRI studies are needed to better understand the pathological processes underlying this novel entity.

Gray Matter Involvement in Radiologically Isolated Syndrome.

The unanticipated magnetic resonance imaging (MRI) detection in the brain of asymptomatic subjects of white matter lesions suggestive of multiple sclerosis has recently been named as radiologically isolated syndrome (RIS). The pathophysiological processes of RIS remain largely unknown and questions as to whether gray matter alterations actually occur in this entity are yet to be investigated in more detail. By means of a 3 T multimodal MRI approach, we searched for cortical and deep gray matter changes in a cohort of RIS patients. Seventeen RIS patients, 17 clinically isolated syndrome (CIS) patients (median disease duration from symptom onset = 12 months), and 17 healthy controls underwent MRI and neuropsychological testing. Normalized deep gray matter volumes and regional cortical thickness were assessed using FreeSurfer. SIENAX was used to obtain normalized global and cortical brain volumes. Voxelwise morphometry analysis was performed by using SPM8 software to localize regions of brain tissue showing significant changes of fractional anisotropy or mean diffusivity. Although no differences were observed between CIS and healthy controls groups, RIS patients showed significantly lower normalized cortical volume (673 ±â€Š27.07 vs 641 ±â€Š35.88 [cm³â€Š× 10³, Tukey P test = 0.009) and mean thalamic volume (0.0051 ±â€Š0.4 vs 0.0046 ±â€Š0.4 mm, P = 0.014) compared with healthy controls. RIS patients also showed significant thinning in a number of cortical areas, that were primarily distributed in frontal and temporal lobes (P < 0.05, uncorrected). Strong correlations were observed between T2-white matter lesion volume and regional cortical thickness (rho spearman ranging from 0.60 to 0.80). Our data suggest that white matter lesions on T2-weighted images are not the only hallmark of RIS. Future longitudinal studies with larger samples are warranted to better clarify the effect of RIS-related white matter lesions on gray matter tissue.

ASAP (Automatic Software for ASL Processing): A toolbox for processing Arterial Spin Labeling images.

The method of Arterial Spin Labeling (ASL) has experienced a significant rise in its application to functional imaging, since it is the only technique capable of measuring blood perfusion in a truly non-invasive manner. Currently, there are no commercial packages for processing ASL data and there is no recognized standard for normalizing ASL data to a common frame of reference. This work describes a new Automated Software for ASL Processing (ASAP) that can automatically process several ASL datasets. ASAP includes functions for all stages of image pre-processing: quantification, skull-stripping, co-registration, partial volume correction and normalization. To assess the applicability and validity of the toolbox, this work shows its application in the study of hypoperfusion in a sample of healthy subjects at risk of progressing to Alzheimer's disease. ASAP requires limited user intervention, minimizing the possibility of random and systematic errors, and produces cerebral blood flow maps that are ready for statistical group analysis. The software is easy to operate and results in excellent quality of spatial normalization. The results found in this evaluation study are consistent with previous studies that find decreased perfusion in Alzheimer's patients in similar regions and demonstrate the applicability of ASAP.

Effect of Water T2 Shortening in the Quantification of in-vitro Proton MR Spectroscopy.

PURPOSE: This work studies the relationship between in-vitro Proton Magnetic Resonance Spectroscopy metabolite quantification and water T2 decay. MATERIALS AND METHODS: An in-vitro correspondence is established between the iron accumulation and the shortening of water T2 relaxation times using seven spherical phantoms, 6 of them were doped with an increasing concentration of iron metal nanoparticles solution. This is later proposed as a source of error during the LCModel metabolite quantification of either absolute concentrations or ratios. RESULTS: The Pearson's correlation coefficient between water T2 values against absolute metabolite concentrations was on average [r] = 0.97 and on average [r] = 0.85 for metabolite ratios. CONCLUSION: These results suggest that the shortening of T2 values should be taken into account when performing metabolite quantification. Also, the need of demonstrated similar results in in-vivo studies, since the presence of iron deposits or other factors affecting the water T2 decay measurements could explain part of the inter-subject variability in the metabolite concentration and ratio quantification.

The partial volume effect in the quantification of 1H magnetic resonance spectroscopy in Alzheimer's disease and aging.

1H-MRS variability increases due to normal aging and also as a result of atrophy in grey and white matter caused by neurodegeneration. In this work, an automatic process was developed to integrate data from spectra and high-resolution anatomical images to quantify metabolites, taking into account tissue partial volumes within the voxel of interest avoiding additional spectra acquisitions required for partial volume correction. To evaluate this method, we use a cohort of 135 subjects (47 male and 88 female, aged between 57 and 99 years) classified into 4 groups: 38 healthy participants, 20 amnesic mild cognitive impairment patients, 22 multi-domain mild cognitive impairment patients, and 55 Alzheimer's disease patients. Our findings suggest that knowing the voxel composition of white and grey matter and cerebrospinal fluid is necessary to avoid partial volume variations in a single-voxel study and to decrease part of the variability found in metabolites quantification, particularly in those studies involving elder patients and neurodegenerative diseases. The proposed method facilitates the use of 1H-MRS techniques in statistical studies in Alzheimer's disease, because it provides more accurate quantitative measurements, reduces the inter-subject variability, and improves statistical results when performing group comparisons.

Higher glutamate to glutamine ratios in occipital regions in women with migraine during the interictal state.

BACKGROUND: Glutamate (Glu) and glutamine (Gln) are strongly compartmentalized (in neurons for Glu and in astrocytes for Gln). The visual cortex is the brain region with a higher neuron/astrocyte ratio (the highest neuronal density and the relatively lowest density of astrocytes). Elevations in extracellular Glu or potassium above certain thresholds are likely candidates to be the final common steps in the multiple distinct processes that can lead to cortical spreading depression. Astrocytes play a key role in this phenomenon, by acting as a sink for extracellular Glu and potassium, as well as generally acting as a buffer for the ionic and neurochemical changes that initiate and propagate cortical spreading depression. OBJECTIVE: The purpose of this study was to quantify Glu and Gln to generate Glu/Gln ratios in women with migraine during the interictal state compared with healthy control women. METHODS: Twenty-seven patients with migraine (8 with aura and 19 without aura) and 19 matched healthy controls were included in the study. We performed proton magnetic resonance spectroscopy in the anterior paracingulate cortex and occipital cortex (OC). Spectral analysis was performed by LCModel, allowing a separation of Glu and Gln using a 3T machine. RESULTS: The main result was a significantly higher Glu/Gln ratio in the OC of migraine patients compared with healthy control subjects (4.87 for migraineurs [standard deviation (SD) = 2.74] and 3.42 for controls [SD = 1.52], P = .042). We also observed higher Glu levels (6.98 for migraineurs [SD = 0.85] and 6.22 for controls [SD = 0.97], P = .007) and Glu/creatine + phosphocreatine ratio (1.18 for migraineurs [SD = 0.18] and 1.00 for controls [SD = 0.16], P = .001) in anterior paracingulate cortex in migraine patients but saw no differences in Glu/Gln ratio (2.79 for migraineurs [SD = 1.11] and 2.63 for controls [SD = 1.61], P = .68). CONCLUSION: These findings are consistent with glutamatergic differences in migraine patients during the interictal period compared with healthy controls. We hypothesize that an increased Glu/Gln ratio could arise from neuronal-glial coupling of glutamatergic metabolism differences or an increased neuron/astrocyte ratio in the OC.