Abstract WP141: Prediction of Subclinical Language Deficit Using Machine Learning Based on Post-stroke Functional Connectivity Derived From Low Frequency Oscillations

Stroke causes alterations in local spontaneous neuronal activity and related networks functional connectivity. We hypothesized that these changes occur in patients with post-stroke cognitive impairment (PSCI). Fractional amplitude of low-frequency fluctuations (fALFF) was calculated in 36 patients with cognitive impairment, including 16 patients with hemorrhagic stroke (hPSCI group), 20 patients with ischemic stroke (iPSCI group). Twenty healthy volunteers closely matched to the patient groups with respect to age and gender were selected as the healthy control group (HC group). Regions with significant alteration were regarded as regions of interest (ROIs) using the one-way analysis of variance, and then the seed-based functional connectivity (FC) with other regions in the brain was analyzed. Pearson correlation analyses were performed to investigate the correlation between functional indexes and cognitive performance in patients with PSCI. Our results showed that fALFF values of bilateral posterior cingulate cortex (PCC)/precuneus and bilateral anterior cingulate cortex in the hPSCI group were lower than those in the HC group. Compared with the HC group, fALFF values were lower in the superior frontal gyrus and basal ganglia in the iPSCI group. Correlation analysis showed that the fALFF value of left PCC was positively correlated with MMSE scores and MoCA scores in hPSCI. Besides, the reduction of seed-based FC values was reported, especially in regions of the default-mode network (DMN) and the salience network (SN). Abnormalities of spontaneous brain activity and functional connectivity are observed in PSCI patients. The decreased fALFF and FC values in DMN of patients with hemorrhagic and SN of patients with ischemic stroke may be the pathological mechanism of cognitive impairment. Besides, we showed how to use fALFF values and functional connectivity maps to specify a target map on the cortical surface for repetitive transcranial magnetic stimulation (rTMS).

Download Full-text

Capturing Neuroplastic Changes after iTBS in Patients with Post-Stroke Aphasia: A Pilot fMRI Study

Brain Sciences ◽

10.3390/brainsci11111451 ◽

2021 ◽

Vol 11 (11) ◽

pp. 1451

Author(s):

Shuo Xu ◽

Qing Yang ◽

Mengye Chen ◽

Panmo Deng ◽

Ren Zhuang ◽

...

Keyword(s):

Functional Connectivity ◽

Primary Motor Cortex ◽

High Efficiency ◽

Parietal Lobe ◽

Low Frequency ◽

Brain Regions ◽

Post Stroke ◽

Before And After ◽

Mri Scans ◽

The Right

Intermittent theta-burst stimulation (iTBS) is a high-efficiency transcranial magnetic stimulation (TMS) paradigm that has been applied to post-stroke aphasia (PSA). However, its efficacy mechanisms have not been clarified. This study aimed to explore the immediate effects of iTBS of the primary motor cortex (M1) of the affected hemisphere, on the functional activities and connectivity of the brains of PSA patients. A total of 16 patients with aphasia after stroke received iTBS with 800 pulses for 300 s. All patients underwent motor, language, and cognitive assessments and resting-state functional MRI scans immediately before and after the iTBS intervention. Regional, seed-based connectivity, and graph-based measures were used to test the immediate functional effects of the iTBS intervention, including the fractional amplitude of low-frequency fluctuation (fALFF), degree centrality (DC), and functional connectivity (FC) of the left M1 area throughout the whole brain. The results showed that after one session of iTBS intervention, the fALFF, DC, and FC values changed significantly in the patients’ brains. Specifically, the DC values were significantly higher in the right middle frontal gyrus and parts of the left parietal lobe (p < 0.05), while fALFF values were significantly lower in the right medial frontal lobe and parts of the left intracalcarine cortex (p < 0.05), and the strength of the functional connectivity between the left M1 area and the left superior frontal gyrus was reduced (p < 0.05). Our findings provided preliminary evidences that the iTBS on the ipsilesional M1 could induce neural activity and functional connectivity changes in the motor, language, and other brain regions in patients with PSA, which may promote neuroplasticity and functional recovery.

Download Full-text

Using Low-Frequency Oscillations to Detect Temporal Lobe Epilepsy with Machine Learning

Brain Connectivity ◽

10.1089/brain.2018.0601 ◽

2019 ◽

Vol 9 (2) ◽

pp. 184-193 ◽

Cited By ~ 5

Author(s):

Gyujoon Hwang ◽

Veena A. Nair ◽

Jed Mathis ◽

Cole J. Cook ◽

Rosaleena Mohanty ◽

...

Keyword(s):

Machine Learning ◽

Temporal Lobe Epilepsy ◽

Temporal Lobe ◽

Low Frequency ◽

Low Frequency Oscillations

Download Full-text

Machine Learning-Based Identification of Suicidal Risk in Patients With Schizophrenia Using Multi-Level Resting-State fMRI Features

Frontiers in Neuroscience ◽

10.3389/fnins.2020.605697 ◽

2021 ◽

Vol 14 ◽

Author(s):

Bartosz Bohaterewicz ◽

Anna M. Sobczak ◽

Igor Podolak ◽

Bartosz Wójcik ◽

Dagmara Mȩtel ◽

...

Keyword(s):

Magnetic Resonance Imaging ◽

Machine Learning ◽

Magnetic Resonance ◽

Functional Connectivity ◽

Functional Magnetic Resonance Imaging ◽

Resting State ◽

Low Frequency ◽

Suicidal Risk ◽

Functional Magnetic Resonance ◽

Resonance Imaging

BackgroundSome studies suggest that as much as 40% of all causes of death in a group of patients with schizophrenia can be attributed to suicides and compared with the general population, patients with schizophrenia have an 8.5-fold greater suicide risk (SR). There is a vital need for accurate and reliable methods to predict the SR among patients with schizophrenia based on biological measures. However, it is unknown whether the suicidal risk in schizophrenia can be related to alterations in spontaneous brain activity, or if the resting-state functional magnetic resonance imaging (rsfMRI) measures can be used alongside machine learning (ML) algorithms in order to identify patients with SR.MethodsFifty-nine participants including patients with schizophrenia with and without SR as well as age and gender-matched healthy underwent 13 min resting-state functional magnetic resonance imaging. Both static and dynamic indexes of the amplitude of low-frequency fluctuation (ALFF), the fractional amplitude of low-frequency fluctuations (fALFF), regional homogeneity as well as functional connectivity (FC) were calculated and used as an input for five machine learning algorithms: Gradient boosting (GB), LASSO, Logistic Regression (LR), Random Forest and Support Vector Machine.ResultsAll groups revealed different intra-network functional connectivity in ventral DMN and anterior SN. The best performance was reached for the LASSO applied to FC with an accuracy of 70% and AUROC of 0.76 (p < 0.05). Significant classification ability was also reached for GB and LR using fALFF and ALFF measures.ConclusionOur findings suggest that SR in schizophrenia can be seen on the level of DMN and SN functional connectivity alterations. ML algorithms were able to significantly differentiate SR patients. Our results could be useful in developing neuromarkers of SR in schizophrenia based on non-invasive rsfMRI.

Download Full-text

Predicting Vasovagal Responses: A Model-Based and Machine Learning Approach

Frontiers in Neurology ◽

10.3389/fneur.2021.631409 ◽

2021 ◽

Vol 12 ◽

Author(s):

Theodore Raphan ◽

Sergei B. Yakushin

Keyword(s):

Machine Learning ◽

Beta Blockers ◽

Vasovagal Syncope ◽

Low Frequency ◽

Tilt Test ◽

Learning Approach ◽

Clear Understanding ◽

Low Frequency Oscillations ◽

Separating Hyperplane ◽

Machine Learning Approach

Vasovagal syncope (VVS) or neurogenically induced fainting has resulted in falls, fractures, and death. Methods to deal with VVS are to use implanted pacemakers or beta blockers. These are often ineffective because the underlying changes in the cardiovascular system that lead to the syncope are incompletely understood and diagnosis of frequent occurrences of VVS is still based on history and a tilt test, in which subjects are passively tilted from a supine position to 20° from the spatial vertical (to a 70° position) on the tilt table and maintained in that orientation for 10–15 min. Recently, is has been shown that vasovagal responses (VVRs), which are characterized by transient drops in blood pressure (BP), heart rate (HR), and increased amplitude of low frequency oscillations in BP can be induced by sinusoidal galvanic vestibular stimulation (sGVS) and were similar to the low frequency oscillations that presaged VVS in humans. This transient drop in BP and HR of 25 mmHg and 25 beats per minute (bpm), respectively, were considered to be a VVR. Similar thresholds have been used to identify VVR's in human studies as well. However, this arbitrary threshold of identifying a VVR does not give a clear understanding of the identifying features of a VVR nor what triggers a VVR. In this study, we utilized our model of VVR generation together with a machine learning approach to learn a separating hyperplane between normal and VVR patterns. This methodology is proposed as a technique for more broadly identifying the features that trigger a VVR. If a similar feature identification could be associated with VVRs in humans, it potentially could be utilized to identify onset of a VVS, i.e, fainting, in real time.

Download Full-text

Machine Learning based identification of suicidal risk in patients with schizophrenia using multi-level resting state fMRI features

10.31234/osf.io/85fum ◽

2020 ◽

Author(s):

Bartosz Bohaterewicz ◽

Maria Sobczak Anna ◽

Igor Podolak ◽

Bartosz Wójcik ◽

Dagmara Mętel ◽

...

Keyword(s):

Magnetic Resonance Imaging ◽

Machine Learning ◽

Magnetic Resonance ◽

Functional Connectivity ◽

Functional Magnetic Resonance Imaging ◽

Resting State ◽

Low Frequency ◽

Suicidal Risk ◽

Functional Magnetic Resonance ◽

Resonance Imaging

Background: Some studies suggest that as much as 40% of all causes of death in a group of patients with schizophrenia can be attributed to suicides and compared with the general population, patients with schizophrenia have an 8.5-fold greater suicide risk (SR). There is a vital need for accurate and reliable methods to predict the SR among patients with schizophrenia based on biological measures. However, it is unknown whether the suicidal risk in schizophrenia can be related to alterations in spontaneous brain activity, or if the resting-state functional magnetic resonance imaging (rsfMRI) measures can be used alongside machine learning (ML) algorithms in order to identify patients with SR.Methods: Fifty-nine participants including patients with schizophrenia with and without SR as well as age and gender-matched healthy underwent 13 minutes of resting-state functional magnetic resonance imaging. Both static and dynamic indexes of the amplitude of low-frequency fluctuation (ALFF), the fractional amplitude of low-frequency fluctuations (fALFF), regional homogeneity as well as functional connectivity (FC) was calculated and used as an input for five machine learning algorithms: Gradient boosting (GB), LASSO, Logistic Regression (LR), Random Forest and Support Vector Machine. Results: All groups revealed different internetwork functional connectivity in ventral DMN and anterior SN. The best performance was reached for the LASSO applied to FC with an accuracy of 70% and AUROC of 0.76 (p<0.05). Significant classification ability was also reached for GB and LR using fALFF and ALFF measures.Conclusion: Our findings suggest that SR in schizophrenia can be seen on the level of DMN and SN functional connectivity alterations. ML algorithms were able to significantly differentiate SR patients. Our results could be useful in developing neuromarkers of SR in schizophrenia based on non-invasive rsfMRI.

Download Full-text

Predicting Post-Stroke Somatosensory Function from Resting-State Functional Connectivity: A Feasibility Study

Brain Sciences ◽

10.3390/brainsci11111388 ◽

2021 ◽

Vol 11 (11) ◽

pp. 1388

Author(s):

Xiaoyun Liang ◽

Chia-Lin Koh ◽

Chun-Hung Yeh ◽

Peter Goodin ◽

Gemma Lamp ◽

...

Keyword(s):

Machine Learning ◽

Functional Connectivity ◽

Brain Networks ◽

High Order ◽

Machine Learning Techniques ◽

Support Vector ◽

Post Stroke ◽

Study Results ◽

Somatosensory Function ◽

Low Order

Accumulating evidence shows that brain functional deficits may be impacted by damage to remote brain regions. Recent advances in neuroimaging suggest that stroke impairment can be better predicted based on disruption to brain networks rather than from lesion locations or volumes only. Our aim was to explore the feasibility of predicting post-stroke somatosensory function from brain functional connectivity through the application of machine learning techniques. Somatosensory impairment was measured using the Tactile Discrimination Test. Functional connectivity was employed to model the global brain function. Behavioral measures and MRI were collected at the same timepoint. Two machine learning models (linear regression and support vector regression) were chosen to predict somatosensory impairment from disrupted networks. Along with two feature pools (i.e., low-order and high-order functional connectivity, or low-order functional connectivity only) engineered, four predictive models were built and evaluated in the present study. Forty-three chronic stroke survivors participated this study. Results showed that the regression model employing both low-order and high-order functional connectivity can predict outcomes based on correlation coefficient of r = 0.54 (p = 0.0002). A machine learning predictive approach, involving high- and low-order modelling, is feasible for the prediction of residual somatosensory function in stroke patients using functional brain networks.

Download Full-text