ABSTRACT
BACKGROUND: Low back pain (LBP) is a major public health issue that influences physical and emotional factors integral to the limbic system. This study aims to investigate the association between LBP and brain morphometry alterations as the duration of LBP increases (acute vs. chronic). METHODS: We used the UK Biobank data to investigate the morphological features of the limbic system in acute LBP (N = 115), chronic LBP (N = 243) and controls (N = 358), and tried to replicate our findings with an independent dataset composed of 45 acute LBP participants evaluated at different timepoints throughout 1 year from the OpenPain database. RESULTS: We found that in comparison with chronic LBP and pain-free controls, acute LBP was associated with increased volumes of the nucleus accumbens, amygdala, hippocampus, and thalamus, and increased grey matter volumes in the hippocampus and posterior cingulate gyrus. In the replication cohort, we found non-significantly larger hippocampus and thalamus volumes in the 3-month visit (acute LBP) compared to the 1-year visit (chronic LBP), with similar effect sizes as the UK Biobank dataset. CONCLUSIONS: Our results suggest that acute LBP is associated with dramatic morphometric increases in the limbic system and mesolimbic pathway, which may reflect an active brain response and self-regulation in the early stage of LBP. SIGNIFICANCE: Our study suggests that LBP in the acute phase is associated with the brain morphometric changes (increase) in some limbic areas, indicating that the acute phase of LBP may represent a crucial stage of self-regulation and active response to the disease's onset.
Subject(s)
Acute Pain , Chronic Pain , Low Back Pain , Humans , Low Back Pain/diagnostic imaging , Low Back Pain/psychology , UK Biobank , Biological Specimen Banks , Limbic System/diagnostic imaging , BrainABSTRACT
Literature suggests that attention is a critical cognitive process for pain perception and modulation and may play an important role in placebo and nocebo effects. Here, we investigated how repeated transcranial direct current stimulation (tDCS) applied at the dorsolateral prefrontal cortex (DLPFC) for three consecutive days can modulate the brain functional connectivity (FC) of two networks involved in cognitive control: the frontoparietal network (FPN) and dorsal attention network (DAN), and its association with placebo and nocebo effects. 81 healthy subjects were randomized to three groups: anodal, cathodal, and sham tDCS. Resting state fMRI scans were acquired pre- and post- tDCS on the first and third day of tDCS. An Independent Component Analysis (ICA) was performed to identify the FPN and DAN. ANCOVA was applied for group analysis. Compared to sham tDCS, 1) both cathodal and anodal tDCS increased the FC between the DAN and right parietal operculum; cathodal tDCS also increased the FC between the DAN and right postcentral gyrus; 2) anodal tDCS led to an increased FC between the FPN and right parietal operculum, while cathodal tDCS was associated with increased FC between the FPN and left superior parietal lobule/precuneus; 3) the FC increase between the DAN and right parietal operculum was significantly correlated to the placebo analgesia effect in the cathodal group. Our findings suggest that both repeated cathodal and anodal tDCS could modulate the FC of two important cognitive brain networks (DAN and FPN), which may modulate placebo / nocebo effects.
Subject(s)
Transcranial Direct Current Stimulation , Humans , Nocebo Effect , Prefrontal Cortex/physiology , Brain , PainABSTRACT
Both acupuncture and imagery have shown potential for chronic pain management. However, the mechanisms underlying their analgesic effects remain unclear. This study aims to explore the thalamocortical mechanisms underlying acupuncture and video-guided acupuncture imagery treatment (VGAIT), a combination of acupuncture and guided imagery, using the resting-state functional connectivity (rsFC) of three thalamic subdivisions-the ventral posterolateral thalamus (VPL), mediodorsal thalamus (MD), and motor thalamus subregion (Mthal)-associated with somatosensory, limbic, and motor circuity. Twenty-seven healthy individuals participated in a within-subject randomized crossover design study. Results showed that compared to sham acupuncture, real acupuncture altered the rsFC between the thalamus and default mode network (DMN) (i.e., mPFC, PCC, and precuneus), as well as the prefrontal and somatosensory cortex (SI/SII). Compared to the VGAIT control, VGAIT demonstrated greater rsFC between the thalamus and key nodes within the interoceptive network (i.e., anterior insula, ACC, PFC, and SI/SII), as well as the motor and sensory cortices (i.e., M1, SMA, and temporal/occipital cortices). Furthermore, compared to real acupuncture, VGAIT demonstrated increased rsFC between the thalamus (VPL/MD/Mthal) and task-positive network (TPN). Further correlations between differences in rsFC and changes in the heat or pressure pain threshold were also observed. These findings suggest that both acupuncture- and VGAIT-induced analgesia are associated with thalamocortical networks. Elucidating the underlying mechanism of VGAIT and acupuncture may facilitate their development, particularly VGAIT, which may be used as a potential remote-delivered pain management approach.
ABSTRACT
Subcortical structures, such as the hippocampus, amygdala, and nucleus accumbens (NAcc), play crucial roles in human cognitive, memory, and emotional processing, chronic pain pathophysiology, and are implicated in various psychiatric and neurological diseases. Interventions modulating the activities of these deep brain structures hold promise for improving clinical outcomes. Recently, non-invasive brain stimulation (NIBS) has been applied to modulate brain activity and has demonstrated its potential for treating psychiatric and neurological disorders. However, modulating the above deep brain structures using NIBS may be challenging due to the nature of these stimulations. This study attempts to identify brain surface regions as source targets for NIBS to reach these deep brain structures by integrating functional magnetic resonance imaging (fMRI) and diffusion tensor imaging (DTI). We used resting-state functional connectivity (rsFC) and probabilistic tractography (PTG) analysis to identify brain surface stimulation targets that are functionally and structurally connected to the hippocampus, amygdala, and NAcc in 119 healthy participants. Our results showed that the medial prefrontal cortex (mPFC) is functionally and anatomically connected to all three subcortical regions, while the precuneus is connected to the hippocampus and amygdala. The mPFC and precuneus, two key hubs of the default mode network (DMN), as well as other cortical areas distributed at the prefrontal cortex and the parietal, temporal, and occipital lobes, were identified as potential locations for NIBS to modulate the function of these deep structures. The findings may provide new insights into the NIBS target selections for treating psychiatric and neurological disorders and chronic pain.
ABSTRACT
Transcranial direct current stimulation (tDCS) is a promising non-invasive method to modulate brain excitability. The aim of this study was to better understand the cerebral blood flow (CBF) changes during and after repeated tDCS at the right dorsolateral prefrontal cortex (DLPFC) in healthy participants using pulsed continuous arterial spin labeling (pCASL). Elucidating CBF changes associated with repeated tDCS may shed light on the understanding of the mechanisms underlying the therapeutic effects of tDCS. tDCS was applied for three consecutive days for 20 min at 2 mA, and MRI scans were performed on day 1 and 3. During anodal tDCS, increased CBF was detected in the bilateral thalamus on day 1 and 3 (12% on day 1 and of 14% on day 3) and in the insula on day 1 (12%). After anodal tDCS on day 1, increased CBF was detected in the cerebellum and occipital lobe (11.8%), while both cathodal and sham tDCS were associated with increased CBF in the insula (11% and 10%, respectively). Moreover, anodal tDCS led to increased CBF in the lateral prefrontal cortex and midcingulate cortex in comparison to the sham. These findings suggest that tDCS can modulate the CBF and different tDCS modes may lead to different effects.
ABSTRACT
Both musculoskeletal pain and sleep disturbances are major health problems worldwide. Literature suggests that the two are reciprocally related and both may be associated with changes in C-reactive protein (CRP) levels. However, the relationships among musculoskeletal pain, sleep duration, and CRP remain unclear. In this cross-sectional study, we investigated the relationship between acute and chronic musculoskeletal pain, sleep, and inflammation using the data from the initial visit of the UK Biobank. 17,642 individuals with chronic musculoskeletal pain, 11,962 individuals with acute musculoskeletal pain, and 29,604 pain-free controls were included in the analysis. In addition, we validated the findings using data from the second visit assessment of the UK Biobank. We found that 1) chronic pain was associated with higher CRP levels compared to both acute pain and the pain-free controls; 2) chronic pain was associated with a lower sleep score (a measurement of sleep patterns), compared to acute pain and the pain-free controls; and acute pain was associated with lower sleep scores compared to the controls; 3) there was a significant negative association between the sleep score and CRP; 4) CRP may partially mediate the association between chronic pain and decreased sleep score. However, the effect size of the mediation was rather small, and the pathophysiological significance remains uncertain. Further validation is needed. These findings were partly replicated in the UK Biobank second visit assessment cohort with a smaller sample size. Our findings, which are based on the large UK Biobank dataset, support the interplay between musculoskeletal pain, sleep patterns, and the potential mediating role of CRP on this reciprocal relationship.
Subject(s)
Acute Pain , Chronic Pain , Musculoskeletal Pain , Sleep Duration , Humans , Acute Pain/epidemiology , Biological Specimen Banks , C-Reactive Protein/analysis , Chronic Pain/epidemiology , Cross-Sectional Studies , Musculoskeletal Pain/epidemiology , United Kingdom/epidemiology , Datasets as TopicABSTRACT
OBJECTIVES: Transcutaneous auricular vagus nerve stimulation (taVNS) is a promising treatment option for migraines. This study aims to investigate the modulation effects of different taVNS frequencies along the central vagus nerve pathway in migraineurs. MATERIALS AND METHODS: Twenty-four migraineurs were recruited for a single-blind, crossover magnetic resonance imaging (MRI) study. The study consisted of two taVNS MRI scan sessions, in which either 1-Hz or 20-Hz taVNS was applied in a random order. Seed-based static and dynamic functional connectivity (FC) analyses were performed using two key nodes of the vagus nerve pathway, the nucleus tractus solitarius (NTS) and the locus coeruleus (LC). RESULTS: Static FC (sFC) analysis showed that 1) continuous 1-Hz taVNS resulted in an increase of NTS/LC-occipital cortex sFC and a decrease of NTS-thalamus sFC compared with the pre-1-Hz taVNS resting state, 2) continuous 20-Hz taVNS resulted in an increase of the LC-anterior cingulate cortex (ACC) sFC compared with the pre-20-Hz taVNS resting state, 3) 1-Hz taVNS produced a greater LC-precuneus and LC-inferior parietal cortex sFC than 20 Hz, and 4) 20-Hz taVNS increased LC-ACC and LC-super temporal gyrus/insula sFC in comparison with 1 Hz. Dynamic FC (dFC) analysis showed that compared with the pre-taVNS resting state, 1-Hz taVNS decreased NTS-postcentral gyrus dFC (less variability), 20-Hz taVNS decreased dFC of the LC-superior temporal gyrus and the LC-occipital cortex. Finally, a positive correlation was found between the subjects' number of migraine attacks in the past four weeks and the NTS-thalamus sFC during pre-taVNS resting state. CONCLUSIONS: 1-Hz and 20-Hz taVNS may modulate the sFC and dFC of key nodes in the central vagus nerve pathway differently. Our findings highlight the importance of stimulation parameters (frequencies) in taVNS treatment.
Subject(s)
Migraine Disorders , Vagus Nerve Stimulation , Humans , Magnetic Resonance Imaging/methods , Migraine Disorders/diagnostic imaging , Migraine Disorders/therapy , Single-Blind Method , Vagus Nerve/physiology , Vagus Nerve Stimulation/methods , Cross-Over StudiesABSTRACT
BACKGROUND AND AIMS: This paper aimed to investigate the usefulness of applying machine learning on resting-state fMRI connectivity data to recognize the pattern of functional changes in essential tremor (ET), a disease characterized by slight brain abnormalities, often difficult to detect using univariate analysis. METHODS: We trained a support vector machine with a radial kernel on the mean signals extracted by 14 brain networks obtained from resting-state fMRI scans of 18 ET and 19 healthy control (CTRL) subjects. Classification performance between pathological and control subjects was evaluated using a tenfold cross-validation. Recursive feature elimination was performed to rank the importance of the extracted features. Moreover, univariate analysis using Mann-Whitney U test was also performed. RESULTS: The machine learning algorithm achieved an AUC of 0.75, with four networks (language, primary visual, cerebellum, and attention), which have an essential role in ET pathophysiology, being selected as the most important features for classification. By contrast, the univariate analysis was not able to find significant results among these two conditions. CONCLUSION: The machine learning approach identifies the changes in functional connectivity of ET patients, representing a promising instrument to discriminate specific pathological conditions and find novel functional biomarkers in resting-state fMRI studies.
Subject(s)
Essential Tremor , Humans , Essential Tremor/pathology , Brain , Machine Learning , Cerebellum/diagnostic imaging , Recognition, Psychology , Magnetic Resonance Imaging/methodsABSTRACT
BACKGROUND AND AIMS: To explore the cognitive functioning of ET patients without dementia and delineate its imaging counterpart. METHODS: We enrolled 99 subjects (49 non-demented ET patients and 50 education-matched healthy controls) that underwent neuropsychological and MRI evaluation. In order to identify the cognitive parameters that better reflect the profile of ET patients, we used a double statistical approach: (i) direct comparison between groups and (ii) machine learning approach with feature selection. Then, to evaluate the correlation between cognitive performances and the degree of brain atrophy in the ET group, we included the results derived from the uni- and multivariate analysis in whole-brain voxel-based morphometry (VBM) model. RESULTS: In ET patients, the univariate analysis showed differences in cognitive tests evaluating executive functions (FAB, MCST-CA), verbal memory-delayed recall (RAVLT-DR), and working memory (Digit Span B). The relative scores were significantly worse compared to controls, although within the normal range (subclinical dysfunctions). The machine learning approach also provided similar findings: tests exploring the executive functions, verbal memory, and language (RAVLT-DR, FAB, COWAT, RAVLT-IR, TOKEN) showed the highest importance rank in classification's task. Regardless of the explored test, the MRI analysis revealed a correlation (p < 0.005 uncorrected, whole brain) between test scores and widespread areas including cerebellum, inferior and middle frontal cortices, cingulate cortices, and temporal cortex. CONCLUSION: This study improves the knowledge on cognitive impairment in ET, as our findings demonstrate a heterogeneous pattern of cognitive dysfunction involving memory, executive function, and language domains in the ET group. This clinical profile relates with the deep involvement of the cerebellum and its connections with large-scale brain structures, suggesting that changes spreading in wide-ranging brain pathways may contribute to the physiopathology of cognitive dysfunction in ET.
Subject(s)
Cognitive Dysfunction , Dementia , Essential Tremor , Cognition , Dementia/diagnostic imaging , Essential Tremor/pathology , Humans , Magnetic Resonance Imaging/methods , Memory, Short-Term , Neuropsychological TestsABSTRACT
BACKGROUND: Resting-state-fMRI is a technique used to explore the functional brain architecture in term of brain networks and their interactions. However, the robustness of Resting-state-fMRI analysis is negatively affected by physiological noise caused by subject head motion. The aim of our study was to provide new knowledge about the effect of normal aging on the head motion signals. NEW METHOD: For the first time, we proposed a method for evaluating the most sensitive head motion parameters linked to subjects'aging. We enrolled 14-young(9females; mean-age = 28 ± 4.07) and 14-elderly(9females; mean-age = 66 ± 5.19) subjects. Along three axes(X,Y,Z), we extracted six motions parameters which reflected the head's movements to characterize translations(x,y,z) and rotations(angles phi,theta,psi). We performed:1)univariate analysis for comparing the groups and correlation to investigate the relationship between age and movement parameters; 2)Support-Vector-Machine, using bootstrap and calculating the feature importance. RESULTS: Statistical analyses showed significant association between the aging and some motion's parameters(rotation psi; translations y and z). These results were also confirmed by multivariate analysis with Support-Vector-Machine that presented an AUC of 90 %. COMPARISON TO EXISTING METHODS: The proposed method shows that normal aging produces significant increase in head motion parameters, highlighting the critical effect of motion on resting data analyses in particular considering psi, y and z movements. To our knowledge and at the present, this represents the first study investigating the accurate characterization of motion parameters in aging. CONCLUSIONS: Our results have a high impact to improve healthy control recruitment and appropriately decreasing the risk of signal distortion, according to the age of enrolled subjects.
Subject(s)
Brain Mapping , Magnetic Resonance Imaging , Adult , Aged , Aging , Artifacts , Brain/diagnostic imaging , Head Movements , Humans , Image Processing, Computer-Assisted , Machine Learning , Middle Aged , Motion , Young AdultABSTRACT
INTRODUCTION: There is growing evidence that a proportion of patients with Essential Tremor (ET) may develop a memory impairment over time. However, no studies have evaluated whether hippocampal damage really occur in ET. This study investigated the macro and micro-structural integrity of the hippocampus in ET subjects using a multimodal MRI approach. METHODS: Neuropsychological and MRI data were acquired from 110 participants (60 patients with ET and 50 age-, sex-, and education-matched healthy controls [HC]). Whole-brain T1-weighted and Diffusion Tensor Imaging (DTI) were performed to assess macro-and microstructural alterations. MRI parameters (volume; mean diffusivity [MD]; fractional anisotropy [FA]) of bilateral hippocampi were obtained. In order to evaluate the relationship between MRI alterations and neurocognitive impairment, hippocampal parameters were also correlated with cognitive test scores. RESULTS: Compared to controls, ET patients showed a subclinical memory impairment with significantly lower memory scores, but within the normal ranges. Despite the subclinical damage, however, ET patients showed a significant increase in MD values in the bilateral hippocampi in comparison with HC. A significant correlation was also found between MD and memory scores in ET. CONCLUSION: This study improves the knowledge on memory impairment in ET, as our results demonstrate for the first time the hippocampal microstructural damage related to subclinical memory impairment in ET patients. Further studies are needed before these findings can be considered predictive of a distinct ET subtype or suggestive of a co-occurent dementia.
Subject(s)
Essential Tremor/pathology , Hippocampus/pathology , Memory Disorders/physiopathology , Aged , Diffusion Tensor Imaging , Essential Tremor/complications , Essential Tremor/diagnosis , Female , Hippocampus/diagnostic imaging , Humans , Magnetic Resonance Imaging , Male , Memory Disorders/etiology , Middle AgedABSTRACT
The intricate relationships between innate immunity and brain diseases raise increased interest across the wide spectrum of neurodegenerative and neuropsychiatric disorders. Barriers, such as the blood-brain barrier, and innate immunity cells such as microglia, astrocytes, macrophages, and mast cells are involved in triggering disease events in these groups, through the action of many different cytokines. Chronic inflammation can lead to dysfunctions in large-scale brain networks. Neurodegenerative diseases, such as Alzheimer's disease, Parkinson's disease, Huntington's disease, amyotrophic lateral sclerosis, and frontotemporal dementia, are associated with a substrate of dysregulated immune responses that impair the central nervous system balance. Recent evidence suggests that similar phenomena are involved in psychiatric diseases, such as depression, schizophrenia, autism spectrum disorders, and post-traumatic stress disorder. The present review summarizes and discusses the main evidence linking the innate immunological response in neurodegenerative and psychiatric diseases, thus providing insights into how the responses of innate immunity represent a common denominator between diseases belonging to the neurological and psychiatric sphere. Improved knowledge of such immunological aspects could provide the framework for the future development of new diagnostic and therapeutic approaches.
Subject(s)
Immunity, Innate , Mental Disorders/immunology , Neurodegenerative Diseases/immunology , Animals , HumansABSTRACT
Importance: Interictal epileptiform discharges (IEDs) in electroencephalograms (EEGs) are a biomarker of epilepsy, seizure risk, and clinical decline. However, there is a scarcity of experts qualified to interpret EEG results. Prior attempts to automate IED detection have been limited by small samples and have not demonstrated expert-level performance. There is a need for a validated automated method to detect IEDs with expert-level reliability. Objective: To develop and validate a computer algorithm with the ability to identify IEDs as reliably as experts and classify an EEG recording as containing IEDs vs no IEDs. Design, Setting, and Participants: A total of 9571 scalp EEG records with and without IEDs were used to train a deep neural network (SpikeNet) to perform IED detection. Independent training and testing data sets were generated from 13â¯262 IED candidates, independently annotated by 8 fellowship-trained clinical neurophysiologists, and 8520 EEG records containing no IEDs based on clinical EEG reports. Using the estimated spike probability, a classifier designating the whole EEG recording as positive or negative was also built. Main Outcomes and Measures: SpikeNet accuracy, sensitivity, and specificity compared with fellowship-trained neurophysiology experts for identifying IEDs and classifying EEGs as positive or negative or negative for IEDs. Statistical performance was assessed via calibration error and area under the receiver operating characteristic curve (AUC). All performance statistics were estimated using 10-fold cross-validation. Results: SpikeNet surpassed both expert interpretation and an industry standard commercial IED detector, based on calibration error (SpikeNet, 0.041; 95% CI, 0.033-0.049; vs industry standard, 0.066; 95% CI, 0.060-0.078; vs experts, mean, 0.183; range, 0.081-0.364) and binary classification performance based on AUC (SpikeNet, 0.980; 95% CI, 0.977-0.984; vs industry standard, 0.882; 95% CI, 0.872-0.893). Whole EEG classification had a mean calibration error of 0.126 (range, 0.109-0.1444) vs experts (mean, 0.197; range, 0.099-0.372) and AUC of 0.847 (95% CI, 0.830-0.865). Conclusions and Relevance: In this study, SpikeNet automatically detected IEDs and classified whole EEGs as IED-positive or IED-negative. This may be the first time an algorithm has been shown to exceed expert performance for IED detection in a representative sample of EEGs and may thus be a valuable tool for expedited review of EEGs.
Subject(s)
Electroencephalography , Epilepsy/diagnosis , Neural Networks, Computer , Signal Processing, Computer-Assisted , Software , Humans , Sensitivity and SpecificityABSTRACT
Importance: The validity of using electroencephalograms (EEGs) to diagnose epilepsy requires reliable detection of interictal epileptiform discharges (IEDs). Prior interrater reliability (IRR) studies are limited by small samples and selection bias. Objective: To assess the reliability of experts in detecting IEDs in routine EEGs. Design, Setting, and Participants: This prospective analysis conducted in 2 phases included as participants physicians with at least 1 year of subspecialty training in clinical neurophysiology. In phase 1, 9 experts independently identified candidate IEDs in 991 EEGs (1 expert per EEG) reported in the medical record to contain at least 1 IED, yielding 87â¯636 candidate IEDs. In phase 2, the candidate IEDs were clustered into groups with distinct morphological features, yielding 12â¯602 clusters, and a representative candidate IED was selected from each cluster. We added 660 waveforms (11 random samples each from 60 randomly selected EEGs reported as being free of IEDs) as negative controls. Eight experts independently scored all 13â¯262 candidates as IEDs or non-IEDs. The 1051 EEGs in the study were recorded at the Massachusetts General Hospital between 2012 and 2016. Main Outcomes and Measures: Primary outcome measures were percentage of agreement (PA) and beyond-chance agreement (Gwet κ) for individual IEDs (IED-wise IRR) and for whether an EEG contained any IEDs (EEG-wise IRR). Secondary outcomes were the correlations between numbers of IEDs marked by experts across cases, calibration of expert scoring to group consensus, and receiver operating characteristic analysis of how well multivariate logistic regression models may account for differences in the IED scoring behavior between experts. Results: Among the 1051 EEGs assessed in the study, 540 (51.4%) were those of females and 511 (48.6%) were those of males. In phase 1, 9 experts each marked potential IEDs in a median of 65 (interquartile range [IQR], 28-332) EEGs. The total number of IED candidates marked was 87â¯636. Expert IRR for the 13â¯262 individually annotated IED candidates was fair, with the mean PA being 72.4% (95% CI, 67.0%-77.8%) and mean κ being 48.7% (95% CI, 37.3%-60.1%). The EEG-wise IRR was substantial, with the mean PA being 80.9% (95% CI, 76.2%-85.7%) and mean κ being 69.4% (95% CI, 60.3%-78.5%). A statistical model based on waveform morphological features, when provided with individualized thresholds, explained the median binary scores of all experts with a high degree of accuracy of 80% (range, 73%-88%). Conclusions and Relevance: This study's findings suggest that experts can identify whether EEGs contain IEDs with substantial reliability. Lower reliability regarding individual IEDs may be largely explained by various experts applying different thresholds to a common underlying statistical model.
Subject(s)
Epilepsy/diagnosis , Electroencephalography , Female , Humans , Male , Observer Variation , Prospective Studies , Reproducibility of ResultsABSTRACT
Machine Learning application on clinical data in order to support diagnosis and prognostic evaluation arouses growing interest in scientific community. However, choice of right algorithm to use was fundamental to perform reliable and robust classification. Our study aimed to explore if different kinds of Machine Learning technique could be effective to support early diagnosis of Multiple Sclerosis and which of them presented best performance in distinguishing Multiple Sclerosis patients from control subjects. We selected following algorithms: Random Forest, Support Vector Machine, Naïve-Bayes, K-nearest-neighbor and Artificial Neural Network. We applied the Independent Component Analysis to resting-state functional-MRI sequence to identify brain networks. We found 15 networks, from which we extracted the mean signals used into classification. We performed feature selection tasks in all algorithms to obtain the most important variables. We showed that best discriminant network between controls and early Multiple Sclerosis, was the sensori-motor I, according to early manifestation of motor/sensorial deficits in Multiple Sclerosis. Moreover, in classification performance, Random Forest and Support Vector Machine showed same 5-fold cross-validation accuracies (85.7%) using only this network, resulting to be best approaches. We believe that these findings could represent encouraging step toward the translation to clinical diagnosis and prognosis.
Subject(s)
Connectome/methods , Forecasting/methods , Multiple Sclerosis/diagnostic imaging , Adult , Algorithms , Bayes Theorem , Brain , Cognition , Female , Humans , Machine Learning , Magnetic Resonance Imaging/methods , Male , Middle Aged , Rest , Support Vector MachineABSTRACT
Alzheimer's disease (AD) and Parkinson's disease with dementia (PDD) are characterized by a different mnesic failure, particularly in memory cued recall. Although hippocampal involvement has been shown in both these diseases, it remains unknown whether a selective damage of specific subfields within the hippocampus may be responsible for the peculiar mnesic profile observed in AD and PDD. To explore this topic, we combined a multimodal 3â¯T-MRI hippocampal evaluation (whole-brain T1-weighted and diffusion tensor imaging) with a hippocampal-targeted neuropsychological assessment (Free and Cued Selective Reminding Test [FCSRT]) in 22 AD subjects, 18 PDD and 17 healthy controls. Macro- and microstructural features (volume; shape; mean diffusivity [MD]; fractional anisotropy [FA]) of bilateral hippocampi (whole and subfields) were obtained. Correlations between MRI-derived parameters and neuropsychological evaluations were performed. In the comparison between AD and PDD, the multimodal analysis allowed us to identify that subiculum, CA1 and CA4-DG were differently involved in these diseases and correlated with immediate and delayed total recall items of FCSRT. Moreover, compared to controls, AD showed a reduction in almost all subfields, with a MD increase in the same regions, whereas PDD displayed a volume loss, less severe than AD, more evident in the CA2-3 and presubiculum subfields. Our study provides new evidence that hippocampal subregions had different vulnerability to damage related to AD and PDD. The combination of the in vivo analysis of hippocampal subfields with the FCSRT paradigm provided important insights into whether changes within specific hippocampal subfields are related to the different mnesic profile in AD and PDD patients.