Hakseung Kim

Hack-Jin Lee

Seung-Bo Lee

In-Nea Wang

Jeongjun Park

Jin-Hyung Kim

Se-Ho Lee

Cheol-Hui Lee

Dong-Kyu Kim

Jae-Wook Heo

Min-Kyung Jung

Dong-Young Kim

Jun-Su Park

Joung-Woo Hyung

Seung-Hoon Yang

Kyung Ah Kim

MinJu Kim

Hwa-Yeon Lee

Brain Lab.은 고려대학교 정보대학 뇌공학과 소속 김동주 교수님 연구실로, 뇌공학 분야 세계 최고 수준의 연구를 진행하고 있습니다. 전자약 (Electroceutical), Brain-Computer Interface (BCI), 생체신호 및 의료 영상 기반 의료 인공지능 등 다양한 분야에서 연구를 진행하고 있으며 최상위권 저널에 논문을 게재하고 있습니다.

NeuroXAI: Adaptive, robust, explainable surrogate framework for determination of channel importance in EEG application

Hyperosmolar therapy response in traumatic brain injury: Explainable artificial intelligence based long-term time series forecasting approach

Effects of altered functional connectivity on motor imagery brain–computer interfaces based on the laterality of paralysis in hemiplegia patients

Prediction of Serious Intracranial Hypertension from Low-Resolution Neuromonitoring in Traumatic Brain Injury: An Explainable Machine Learning Approach

SleepExpertNet: high‑performance and class‑balanced deep learning approach inspired from the expert neurologists for sleep stage classification

DeepCNAP: A Deep Learning Approach for Continuous Noninvasive Arterial Blood Pressure Monitoring Using Photoplethysmography

Prediction of Life-Threatening Intracranial Hypertension During the Acute Phase of Traumatic Brain Injury Using Machine Learning

Brain-Controlled Robotic Arm System based on Multi-Directional CNN-BiLSTM Network using EEG Signals

Comparative Analysis of Features Extracted from EEG Spatial, Spectral and Temporal Domains for Binary and Multiclass Motor Imagery Classification

Artifact removal from neurophysiological signals: impact on intracranial and arterial pressure monitoring in traumatic brain injury

Novel index for predicting mortality during the first 24 hours after traumatic brain injury

Changes in the gray and white matter of patients with ischemic-edematous insults after traumatic brain injury

Automated artifact elimination of physiological signals using a deep belief network: An application for continuously measured arterial blood pressure waveforms

Hemodynamic Instability and Cardiovascular Events After Traumatic Brain Injury Predict Outcome After Artifact Removal With Deep Belief Network Analysis

Abilities of a Densitometric Analysis of Computed Tomography Images and Hemorrhagic Parameters to Predict Outcome Favorability in Patients With Intracerebral Hemorrhage

Cerebrovascular pressure reactivity monitoring using wavelet analysis in traumatic brain injury patients: A retrospective study

Finite element analysis for normal pressure hydrocephalus: The effects of the integration of sulci

Quantitative analysis of computed tomography images and early detection of cerebral edema for pediatric traumatic brain injury patients: Retrospective study

Negative-Sample-Free Contrastive Self-Supervised Learning for Electroencephalogram-Based Motor Imagery Classification

Inherent seizure susceptibility in patients with antihistamine-induced acute symptomatic seizure: a resting-state EEG analysis

Autonomic Dysfunction in Sleep Disorders: From Neurobiological Basis to Potential Therapeutic Approaches

Intracranial Densitometry-Augmented Machine Learning Enhances the Prognostic Value of Brain CT in Pediatric Patients With Traumatic Brain Injury: A Retrospective Pilot Study

Multi-Domain Convolutional Neural Networks for Lower-Limb Motor Imagery Using Dry vs. Wet Electrodes

Cerebrospinal fluid dynamics correlate with neurogenic claudication in lumbar spinal stenosis

Lower Breakpoint of Intracranial Amplitude-Pressure Relationship in Normal Pressure Hydrocephalus

Impaired cerebral compensatory reserve is associated with admission imaging characteristics of diffuse insult in traumatic brain injury

Validation of Pressure Reactivity and Pulse Amplitude Indices Against the Lower Limit of Autoregulation, Part I: Experimental Intra-Cranial Hypertension

Validation of Davson’s equation in patients suffering from idiopathic normal pressure hydrocephalus

Mathematical Modelling of CSF Pulsatile Flow in Aqueduct Cerebri

Monitoring of Optimal Cerebral Perfusion Pressure in Traumatic Brain Injured Patients Using a Multi-Window Weighting Algorithm

An Association Between ICP-Derived Data and Outcome in TBI Patients: The Role of Sample Size

Chlorhexidine and silver sulfadiazine coating on central venous catheters is not sufficient for protection against catheter-related infection: Simulation-based laboratory research with clinical validation

Cerebrospinal fluid dynamics at the lumbosacral level in patients with spinal stenosis: A pilot study: CSF DYNAMICS IN SPINAL STENOSIS

Spectral analysis of intracranial pressure: Is it helpful in the assessment of shunt functioning in-vivo?

Finite Element Model for Hydrocephalus and Idiopathic Intracranial Hypertension

Morphological Feature Extraction From a Continuous Intracranial Pressure Pulse via a Peak Clustering Algorithm

Cerebrospinal fluid dynamics at the lumbosacral level in patients with spinal stenosis

Finite element analysis of periventricular lucency in hydrocephalus: extravasation or transependymal CSF absorption?

Cerebral critical closing pressure in hydrocephalus patients undertaking infusion tests

Porohyperelastic anatomical models for hydrocephalus and idiopathic intracranial hypertension

Phase-shift between arterial flow and ICP pulse during infusion test

Cerebral Vasospasm Affects Arterial Critical Closing Pressure

Thresholds of resistance to CSF outflow in predicting shunt responsiveness

How does moderate hypocapnia affect cerebral autoregulation in response to changes in perfusion pressure in TBI patients?

Continuous Monitoring of the Monro-Kellie Doctrine: Is It Possible?

Cerebral arterial compliance in patients with internal carotid artery disease

Effect of Hyper- and Hypocapnia on Cerebral Arterial Compliance in Normal Subjects

Evaluation of the cerebrovascular pressure reactivity index using non-invasive finapres arterial blood pressure

Cerebrospinal compensation of pulsating cerebral blood volume in hydrocephalus

Transient Changes in Brain Tissue Oxygen in Response to Modifications of Cerebral Perfusion Pressure: An Observational Study

What Shapes Pulse Amplitude of Intracranial Pressure?

A comparison study of cerebral autoregulation assessed with transcranial Doppler and cortical laser Doppler flowmetry

The monitoring of relative changes in compartmental compliances of brain

Brain-controlled robotic arm system based on multi-directional CNN-BiLSTM network using EEG signals

What shapes pulse amplitude of intracranial pressure?

Comparative analysis of features extracted from EEG spatial, spectral and temporal domains for binary and multiclass motor imagery classification

Continuous monitoring of the Monro-Kellie doctrine: is it possible?

Index of cerebrospinal compensatory reserve in hydrocephalus

Autonomic dysfunction in sleep disorders: from neurobiological basis to potential therapeutic approaches

Monitoring of optimal cerebral perfusion pressure in traumatic brain injured patients using a multi-window weighting algorithm

Plateau waves in head injured patients requiring neurocritical care

Cerebrovascular pressure reactivity monitoring using wavelet analysis in traumatic brain injury patients: a retrospective study

Deepcnap: A deep learning approach for continuous noninvasive arterial blood pressure monitoring using photoplethysmography

Classification of hand motions within EEG signals for non-invasive BCI-based robot hand control

Validation of pressure reactivity and pulse amplitude indices against the lower limit of autoregulation, part I: experimental intracranial hypertension

Effect of hyper‐and hypocapnia on cerebral arterial compliance in normal subjects

Trajectory decoding of arm reaching movement imageries for brain-controlled robot arm system

Quantitative analysis of computed tomography images and early detection of cerebral edema for pediatric traumatic brain injury patients: retrospective study

An association between ICP-derived data and outcome in TBI patients: the role of sample size

Dong-Joo Kim

Brain-machine interfaces (BMIs) can be used to decode brain activity into commands to control external devices. This paper presents the decoding of intuitive upper extremity imagery for multi-directional arm reaching tasks in three-dimensional (3D) environments. We designed and implemented an experimental environment in which electroencephalogram (EEG) signals can be acquired for movement execution and imagery. Fifteen subjects participated in our experiments. We proposed a multi-directional convolution neural network-bidirectional long short-term memory network (MDCBN)-based deep learning framework. The decoding performances for six directions in 3D space were measured by the correlation coefficient (CC) and the normalized root mean square error (NRMSE) between predicted and baseline velocity profiles. The grand-averaged CCs of multi-direction were 0.47 and 0.45 for the execution and …

The pulsatile component of intracranial pressure (ICP) has been shown to be a predictor of outcome in normal pressure hydrocephalus (NPH) and traumatic brain injury (TBI). Experimental studies have demonstrated that the pulse amplitude of ICP (AMP ICP ) is dependent on the mean ICP (mICP), and on the pulse amplitude of the cerebral arterial blood volume (AMP CaBV ), according to the exponential craniospinal compliance curve. In this study, we compared the influence of mICP and AMP CaBV on AMP ICP in patients with NPH (infusion study) and TBI (spontaneous recording). We retrospectively analyzed 25 NPH and 43 TBI patients with continuous monitoring of ICP and cerebral blood flow velocity (CBFV), as assessed with transcranial doppler. AMP CaBV was extracted from the CBFV waveform. The influence of mICP and AMP CaBV on AMP ICP were determined using partial coefficients a, b, and c of the …

The electroencephalogram (EEG) remains the predominant source of neurophysiological signals for motor imagery-based brain-computer interfaces (MI-BCIs). Various features can be derived from three distinctive domains (i.e., spatial, temporal and spectral); however, the efficacies of the existing feature extraction methods when discriminating complex multiclass MI tasks have yet to be reported. This study investigates the performances of EEG feature extraction techniques from varying domains against different levels of complex, multiclass MI tasks. Ten healthy volunteers underwent multiple complex MI tasks via a robotic arm (i.e., hand grasping and wrist twisting; grasp, spread, pronation and supination). The discrimination performances of various feature extraction (i.e., common spatial patterns (CSP), time domain parameters (TDP), and power spectral density (PSD)) and classification methods for EEG were …

The Monro-Kellie doctrine describes the principle of homeostatic intracerebral volume regulation, which stipulates that the total volume of the parenchyma, cerebrospinal fluid, and blood remains constant. Hypothetically, a slow shift (e.g., brain edema development) in the irregular vasomotion-driven exchanges of these compartmental volumes may lead to increased intracranial hypertension. To evaluate this paradigm in a clinical setting and measure the processes involved in the regulation of systemic intracranial volume, we quantified cerebral blood flow velocity (CBFv) in the middle cerebral artery, arterial blood pressure (ABP), and intracranial pressure (ICP), in 238 brain-injured subjects. Relative changes in compartmental compliances C a (arterial) and C i (combined venous and CSF compartments) were mathematically estimated using these raw signals through time series analysis; C a and C i were used to …

OBJECTIVE An index of cerebrospinal compensatory reserve (RAP) has been introduced as a potential descriptor of neurological deterioration after head trauma. It is numerically computed as a linear correlation coefficient between the mean intracranial pressure and the pulse amplitude of the pressure waveform. We explore how RAP varies with different forms of physiological or nonphysiological intracranial volume loads in adult hydrocephalus, with and without a functioning cerebrospinal fluid (CSF) shunt. METHODS A database of intracranial pressure recordings during CSF infusion studies and overnight monitoring in hydrocephalic patients was reviewed for clinical comparison of homogeneous subgroups of patients with hypothetical differences of pressure-volume compensatory reserve. The database includes 980 patients of mixed etiology: idiopathic normal pressure …

The study aimed to develop a computational method for assessing relative changes in compartmental compliances within the brain: the arterial bed and the cerebrospinal space. The method utilizes the relationship between pulsatile components in the arterial blood volume, arterial blood pressure (ABP) and intracranial pressure (ICP). It was verified by using clinical recordings of intracranial pressure plateau waves, when massive vasodilatation accompanying plateau waves produces changes in brain compliances of the arterial bed ( C a ) and compliance of the cerebrospinal space ( C i ). Ten patients admitted after head injury with a median Glasgow Coma Score of 6 were studied retrospectively. ABP was directly monitored from the radial artery. Changes in the cerebral arterial blood volume were assessed using Transcranial Doppler (TCD) ultrasonography by digital integration of inflow blood velocity. During …

Sleep disorder has been portrayed as merely a common dissatisfaction with sleep quality and quantity. However, sleep disorder is actually a medical condition characterized by inconsistent sleep patterns that interfere with emotional dynamics, cognitive functioning, and even physical performance. This is consistent with sleep abnormalities being more common in patients with autonomic dysfunction than in the general population. The autonomic nervous system coordinates various visceral functions ranging from respiration to neuroendocrine secretion in order to maintain homeostasis of the body. Because the cell population and efferent signals involved in autonomic regulation are spatially adjacent to those that regulate the sleep-wake system, sleep architecture and autonomic coordination exert effects on each other, suggesting the presence of a bidirectional relationship in addition to shared pathology. The …

Methods to identify an autoregulation guided “optimal” cerebral perfusion pressure (CPPopt) for traumatic brain injury patients (TBI) have been reported through several studies. An important drawback of existing methodology is that CPPopt can be calculated only in ∼50–60% of the monitoring time. In this study, we hypothesized that the CPPopt yield and the continuity can be improved significantly through application of a multi-window and weighting calculation algorithm, without adversely affecting preservation of its prognostic value. Data of 526 severe TBI patients admitted between 2003 and 2015 were studied. The multi-window CPPopt calculation was based on automated curve fitting in pressure reactivity index (PRx)-CPP plots using data from 36 increasing length time windows (2–8 h). The resulting matrix of CPPopts was then averaged in a weighted manner. The yield, continuity, and stability of CPPopt …

Object Plateau waves often develop in neurointensive care patients. They are sudden increases in intracranial pressure (ICP) that lead to dramatic decreases of cerebral perfusion pressure (CPP) and can therefore contribute to ischemic secondary brain insult. The aim of this study was to analyze the occurrence of plateau waves in head injured patients requiring neurocritical care, their relation with cerebral autoregulation and impact on outcome. Methods Data were analyzed retrospectively in 444 head injured patients admitted to Neuroscience Critical Care Unit of Addenbrooke’s Hospital in Cambridge, UK. Arterial blood pressure (ABP), intracranial pressure (ICP), heart rate (HR) were digitally recorded and derived indices calculated. Primary monitoring data, autoregulation indices, outcome of patients, initial CT findings (in a subgroup of patients), brain tissue monitoring data (in a subgroup) were compared …

Background After traumatic brain injury (TBI), the ability of cerebral vessels to appropriately react to changes in arterial blood pressure (pressure reactivity) is impaired, leaving patients vulnerable to cerebral hypo- or hyperperfusion. Although, the traditional pressure reactivity index (PRx) has demonstrated that impaired pressure reactivity is associated with poor patient outcome, PRx is sometimes erratic and may not be reliable in various clinical circumstances. Here, we introduce a more robust transform-based wavelet pressure reactivity index (wPRx) and compare its performance with the widely used traditional PRx across 3 areas: its stability and reliability in time, its ability to give an optimal cerebral perfusion pressure (CPPopt) recommendation, and its relationship with patient outcome. Methods and findings Five hundred and fifteen patients with TBI admitted in Addenbrooke’s Hospital, United Kingdom (March 23rd, 2003 through December 9th, 2014), with continuous monitoring of arterial blood pressure (ABP) and intracranial pressure (ICP), were retrospectively analyzed to calculate the traditional PRx and a novel wavelet transform-based wPRx. wPRx was calculated by taking the cosine of the wavelet transform phase-shift between ABP and ICP. A time trend of CPPopt was calculated using an automated curve-fitting method that determined the cerebral perfusion pressure (CPP) at which the pressure reactivity (PRx or wPRx) was most efficient (CPPopt_PRx and CPPopt_wPRx, respectively). There was a significantly positive relationship between PRx and wPRx (r = 0.73), and wavelet wPRx was more reliable in time (ratio of between-hour …

Arterial blood pressure (ABP) monitoring may permit the early diagnosis and management of cardiovascular disease (CVD); however, existing methods for measuring ABP outside the clinic use inconvenient cuff sphygmomanometry, or do not estimate continuous ABP waveforms. This study proposes a novel deep learning model DeepCNAP for estimating continuous BP waveform from a noninvasively measured photoplethysmography (PPG) signal in real-time. DeepCNAP was designed through the combination of deep convolutional networks and self-attention. The proposed method was constructed via 10-fold cross-validation based on the MIMIC database (the number of subjects = 942, recording time = 374.43 hours). The performance of DeepCNAP was evaluated from two perspectives: estimating ABP from PPG and classifying hemodynamically unstable events (i.e., hypertension, prehypertension, hypotension …

The development of brain-computer interface (BCI) systems that are based on electroencephalography (EEG), and driven by spontaneous movement intentions, is useful for rehabilitation and external device control. In this study, we analyzed the decoding of five different hand executions and imageries from EEG signals, for a robot hand control. Five healthy subjects participated in this experiment. They executed and imagined five sustained hand motions. In this motor execution (ME) and motor imagery (MI) experiment, we proposed a subject-specific time interval selection method, and we used common spatial patterns (CSP) and the regularized linear discriminant analysis (RLDA) for the data analysis. As a result, we classified the five different hand motions offline and obtained average classification accuracies of 56.83% for ME, and 51.01% for MI, respectively. Both results were higher than the obtained accuracies …

The purpose of this study was to provide validation of intracranial pressure (ICP) derived continuous indices of cerebrovascular reactivity against the lower limit of autoregulation (LLA). Utilizing an intracranial hypertension model within white New Zealand rabbits, ICP, transcranial Doppler (TCD), laser Doppler flowmetry (LDF), and arterial blood pressure were recorded. Data were retrospectively analyzed in a cohort of 12 rabbits with adequate signals for interrogating the LLA. We derived continuous indices of cerebrovascular reactivity: PRx (correlation between ICP and mean arterial pressure [MAP]), PAx (correlation between pulse amplitude of ICP [AMP] and MAP), and Lx (correlation between LDF-based cerebral blood flow [CBF] and cerebral perfusion pressure [CPP]). LLA was derived via piecewise linear regression of CPP versus LDF or CPP versus systolic flow velocity (FVs) plots. We then produced error bar …

Objectives: Disturbances in cerebrospinal fluid (CSF) absorption capacity represent one of the primary causes of normal pressure hydrocephalus (NPH). The resistance to CSF outflow ( R out ) may reflect the severity of this disturbance and has been used as a supplementary parameter to predict shunt responsiveness. However, there is no widely accepted threshold of R out in this scenario. This study aimed to determine the most appropriate threshold through meta-analyses. Methods: A total of four previously reported cutoff values of R out (10, 12, 15, and 18 mmHg/ml/minutes) were selected as potential thresholds. A systematic review of existing studies involving NPH, shunt surgeries, and R out was conducted. The pooled results from retrieved articles were subjected to statistical analyses. Results: A total of nine studies were included in the meta-analyses based on the available preoperative R out and postoperative …

ABSTRACT BACKGROUND Changes in partial pressure of carbon dioxide (PaCO2) are associated with a decrease in cerebral blood flow (CBF) during hypocapnia and an increase in CBF during hypercapnia. However, the effects of changes in PaCO2 on cerebral arterial compliance (Ca) are unknown. METHODS We assessed the changes in Ca in 20 normal subjects using monitoring of arterial blood pressure (ABP) and cerebral blood flow velocity (CBFV). Cerebral arterial blood volume (CaBV) was extracted from CBFV. Ca was defined as the ratio between the pulse amplitudes of CaBV (AMP CaBV ) and ABP (AMP ABP ). All parameters were recorded during normo‐, hyper‐, and hypocapnia. RESULTS During hypocapnia, Ca was significantly lower than during normocapnia (.10 ± .04 vs. .17 ± .06; P < .001) secondary to a decrease in AMP CaBV (1.3 ± .4 vs. 1.9 ± .5; P < .001) and a concomitant increase in …

Background Continuous assessment of cerebral compensatory reserve is possible using the moving correlation between pulse amplitude of intra-cranial pressure (AMP) and intra-cranial pressure (ICP), called RAP. Little is known about the behavior and associations of this index in adult traumatic brain injury (TBI). The goal of this study is to evaluate the association between admission cerebral imaging findings and RAP over the course of the acute intensive care unit stay. Methods We retrospectively reviewed 358 adult TBI patients admitted to the Addenbrooke’s Hospital, University of Cambridge, from March 2005 to December 2016. Only non-craniectomy patients were studied. Using archived high frequency physiologic signals, RAP was derived and analyzed over the first 48 h and first 10 days of recording in each patient, using grand mean, percentage of time above various thresholds, and integrated area under …

Development of noninvasive brain-machine interface (BMI) systems based on electroencephalography (EEG), driven by spontaneous movement intentions, is a useful tool for controlling external devices or supporting a neuro- rehabilitation. In this study, we present the possibility of brain-controlled robot arm system using arm trajectory decoding. To do that, we first constructed the experimental system that can acquire the EEG data for not only movement execution (ME) task but also movement imagery (MI) tasks. Five subjects participated in our experiments and performed four directional reaching tasks (Left, right, forward, and backward) in the 3D plane. For robust arm trajectory decoding, we propose a subject-dependent deep neural network (DNN) architecture. The decoding model applies the principle of bi-directional long short-term memory (LSTM) network. As a result, we confirmed the decoding performance (r …

Background The purpose of this study was to identify whether the distribution of Hounsfield Unit (HU) values across the intracranial area in computed tomography (CT) images can be used as an effective diagnostic tool for determining the severity of cerebral edema in pediatric traumatic brain injury (TBI) patients. Methods CT images, medical records and radiology reports on 70 pediatric patients were collected. Based on radiology reports and the Marshall classification, the patients were grouped as mild edema patients (n = 37) or severe edema patients (n = 33). Automated quantitative analysis using unenhanced CT images was applied to eliminate artifacts and identify the difference in HU value distribution across the intracranial area between these groups. Results The proportion of pixels with HU =17 to 24 was highly correlated with the existence of severe cerebral edema ( P <0.01). This proportion was also able to …

Background Many demographic and physiological variables have been associated with TBI outcomes. However, with small sample sizes, making spurious inferences is possible. This paper explores the effect of sample sizes on statistical relationships between patient variables (both physiological and demographic) and outcome. Methods Data from head-injured patients with monitored arterial blood pressure, intracranial pressure (ICP) and outcome assessed at 6 months were included in this retrospective analysis. A univariate logistic regression analysis was performed to obtain the odds ratio for unfavorable outcome. Three different dichotomizations between favorable and unfavorable outcomes were considered. A bootstrap method was implemented to estimate the minimum sample sizes needed to obtain reliable association between physiological and demographic variables with outcome. Results In a univariate …

Objectives: We compared autoregulation monitored with cortical laser Doppler flowmetry (LDF) and autoregulation monitored with transcranial Doppler (TCD) in the middle cerebral artery (MCA) to verify the hypothesis that, following brain trauma, cortical vessel autoregulation to intracranial hypertension is different than assessed in the MCA. Methods: Data collected from 29 head injured patients were analysed retrospectively. Arterial blood pressure (ABP), intracranial pressure (ICP), flow velocity (FV) of the MCA and cortical flux (LDF) were monitored. Indices of cortical autoregulation (Lx) and autoregulation of cerebral blood flow in the MCA (Mx) were calculated as a moving correlation coefficient between slow waves of LDF and cerebral perfusion pressure (CPP) (Lx) or FV and CPP (Mx), respectively. Intact autoregulation was indicated by negative values for Lx and Mx; disturbed autoregulation was reflected by …

Professor

Prediction of life-threatening intracranial hypertension during the acute phase of traumatic brain injury using machine learning

Neuronet: A novel hybrid self-supervised learning framework for sleep stage classification using single-channel eeg

Morphological feature extraction from a continuous intracranial pressure pulse via a peak clustering algorithm

Hemodynamic instability and cardiovascular events after traumatic brain injury predict outcome after artifact removal with deep belief network analysis

Artificial intelligence-enhanced neurocritical care for traumatic brain injury: past, present and future

Prediction of serious intracranial hypertension from low-resolution neuromonitoring in traumatic brain injury: an explainable machine learning approach

SleepExpertNet: high-performance and class-balanced deep learning approach inspired from the expert neurologists for sleep stage classification

An ensemble deep learning approach for sleep stage classification via single-channel EEG and EOG

Finite element analysis for normal pressure hydrocephalus: the effects of the integration of sulci

OBJECTIVE Monitoring intracranial and arterial blood pressure (ICP and ABP, respectively) provides crucial information regarding the neurological status of patients with traumatic brain injury (TBI). However, these signals are often heavily affected by artifacts, which may significantly reduce the reliability of the clinical determinations derived from the signals. The goal of this work was to eliminate signal artifacts from continuous ICP and ABP monitoring via deep learning techniques and to assess the changes in the prognostic capacities of clinical parameters after artifact elimination. METHODS The first 24 hours of monitoring ICP and ABP in a total of 309 patients with TBI was retrospectively analyzed. An artifact elimination model for ICP and ABP was constructed via a stacked convolutional autoencoder (SCAE) and convolutional neural network (CNN) with 10-fold cross …

OBJECT Periventricular lucency (PVL) is often observed in the hydrocephalic brain on CT or MRI. Earlier studies have proposed the extravasation of ventricular CSF into the periventricular white matter or transependymal CSF absorption as possible causes of PVL in hydrocephalus. However, there is insufficient evidence for either theory to be conclusive. METHODS A finite element (FE) model of the hydrocephalic brain with detailed anatomical geometry was constructed to investigate the possible mechanism of PVL in hydrocephalus. The initiation of hydrocephalus was modeled by applying a transmantle pressure gradient (TPG). The model was exposed to varying TPGs to investigate the effects of different geometrical characteristics on the distribution of PVL. The edema map was derived based on the interstitial pore pressure. RESULTS The model …

Intracranial hypertension (IH) following acute phase traumatic brain injury (TBI) is associated with high mortality. Objective: This study proposes a novel parameter that may identify a potentially life-threatening IH (LTH) event and designs a machine learning model to predict LTH. Continuous recordings of intracranial pressure (ICP) and arterial blood pressure (ABP) from 273 TBI patients were used as the development dataset. The pressure-time dose (PTD) and pressure reactivity index (PRx) were calculated for each IH event, and an IH event with PRx > 0 and PTD > 5 was considered an LTH event. The association between the LTH parameters accumulated over five days and mortality was analyzed. A categorical boosting (CatBoost) model was employed to predict the occurrence of a future LTH event from the onset of IH using the ABP- and ICP-related parameters. Training and validation were performed on a total …

The classification of sleep stages is a pivotal aspect of diagnosing sleep disorders and evaluating sleep quality. However, the conventional manual scoring process, conducted by clinicians, is time-consuming and prone to human bias. Recent advancements in deep learning have substantially propelled the automation of sleep stage classification. Nevertheless, challenges persist, including the need for large datasets with labels and the inherent biases in human-generated annotations. This paper introduces NeuroNet, a self-supervised learning (SSL) framework designed to effectively harness unlabeled single-channel sleep electroencephalogram (EEG) signals by integrating contrastive learning tasks and masked prediction tasks. NeuroNet demonstrates superior performance over existing SSL methodologies through extensive experimentation conducted across three polysomnography (PSG) datasets. Additionally, this study proposes a Mamba-based temporal context module to capture the relationships among diverse EEG epochs. Combining NeuroNet with the Mamba-based temporal context module has demonstrated the capability to achieve, or even surpass, the performance of the latest supervised learning methodologies, even with a limited amount of labeled data. This study is expected to establish a new benchmark in sleep stage classification, promising to guide future research and applications in the field of sleep analysis.

Objective An increase in intracranial pressure (ICP) is frequently observed in patients with severe traumatic brain injury (TBI). The information derived from the observation of temporal changes in the mean ICP is insufficient for assessment of the compensatory reserve of the injured brain. This assessment can be achieved via continuous morphological analysis of the pulse waveform of the ICP. Methods Continuous arterial blood pressure (ABP) and ICP recordings from 292 TBI patients were analyzed. The algorithm extracted morphological landmarks (peaks, troughs, and flats) from the ICP. Among the extracted peaks, P1, P2, and P3 were assigned through peak clustering. The performance of the proposed method was validated through a comparison of the algorithm-defined peaks and those manually identified by experienced observers. Results The proposed algorithm successfully identified the three …

Artifacts in physiological signals acquired during intensive care have the potential to be recognized as critical pathological events and lead to misdiagnosis or mismanagement. Manual artifact removal necessitates significant labor-time intensity and is subject to inter- and intra-observer variability. Various methods have been proposed to automate the task; however, the methods are yet to be validated, possibly due to the diversity of artifact types. Deep belief networks (DBNs) have been shown to be capable of learning generative and discriminative feature extraction models, hence suitable for classifying signals with multiple features. This study proposed a DBN-based model for artifact elimination in pulse waveform signals, which incorporates pulse segmentation, pressure normalization and decision models using DBN, and applied the model to artifact removal in monitoring arterial blood pressure (ABP). When …

Background: Hemodynamic instability and cardiovascular events heavily affect the prognosis of traumatic brain injury. Physiological signals are monitored to detect these events. However, the signals are often riddled with faulty readings, which jeopardize the reliability of the clinical parameters obtained from the signals. A machine-learning model for the elimination of artifactual events shows promising results for improving signal quality. However, the actual impact of the improvements on the performance of the clinical parameters after the elimination of the artifacts is not well studied. Materials and Methods: The arterial blood pressure of 99 subjects with traumatic brain injury was continuously measured for 5 consecutive days, beginning on the day of admission. The machine-learning deep belief network was constructed to automatically identify and remove false incidences of hypotension, hypertension, bradycardia …

In neurointensive care units (NICUs), particularly in cases involving traumatic brain injury (TBI), swift and accurate decision-making is critical because of rapidly changing patient conditions and the risk of secondary brain injury. The use of artificial intelligence (AI) in NICU can enhance clinical decision support and provide valuable assistance in these complex scenarios. This article aims to provide a comprehensive review of the current status and future prospects of AI utilization in the NICU, along with the challenges that must be overcome to realize this. Presently, the primary application of AI in NICU is outcome prediction through the analysis of preadmission and high-resolution data during admission. Recent applications include augmented neuromonitoring via signal quality control and real-time event prediction. In addition, AI can integrate data gathered from various measures and support minimally invasive neuromonitoring to increase patient safety. However, despite the recent surge in AI adoption within the NICU, the majority of AI applications have been limited to simple classification tasks, thus leaving the true potential of AI largely untapped. Emerging AI technologies, such as generalist medical AI and digital twins, harbor immense potential for enhancing advanced neurocritical care through broader AI applications. If challenges such as acquiring high-quality data and ethical issues are overcome, these new AI technologies can be clinically utilized in the actual NICU environment. Emphasizing the need for continuous research and development to maximize the potential of AI in the NICU, we anticipate that this will further enhance the …

OBJECTIVE Failure of cerebral autoregulation and subsequent hypoperfusion is common during the acute phase of traumatic brain injury (TBI). The cerebrovascular pressure-reactivity index (PRx) indirectly reflects cerebral autoregulation and has been used to derive optimal cerebral perfusion pressure (CPP). This study provides a method for the use of a combination of PRx, CPP, and intracranial pressure (ICP) to better evaluate the extent of cerebral hypoperfusion during the first 24 hours after TBI, allowing for a more accurate prediction of mortality risk. METHODS Continuous ICP and arterial blood pressure (ABP) signals acquired from 295 TBI patients during the first 24 hours after admission were retrospectively analyzed. The CPP at the lowest PRx was determined as the optimal CPP (CPPopt). The duration of a severe hypoperfusion event (dHP) was defined as the …

OBJECT Brain deformation can be seen in hydrocephalus and idiopathic intracranial hypertension (IIH) via medical images. The phenomenology of local effects, brain shift, and raised intracranial pressure and herniation are textbook concepts. However, there are still uncertainties regarding the specific processes that occur when brain tissue is subject to the mechanical stress of different temporal and spatial profiles of the 2 neurological disorders. Moreover, recent studies suggest that IIH and hydrocephalus may be diseases with opposite pathogenesis. Nevertheless, the similarities and differences between the 2 subjects have not been thoroughly investigated. METHODS An anatomical porohyperelastic finite element model was used to assess the brain tissue responses associated with hydrocephalus and IIH. The same set of boundary conditions, with the exception of brain …

There is a strong association between intracranial hypertension (IH) that occurs following the acute phase of traumatic brain injury (TBI) and negative outcomes. This study proposes a pressure–time dose (PTD)-based parameter that may specify a possible serious IH (SIH) event and develops a model to predict SIH. The minute-by-minute signals of arterial blood pressure (ABP) and intracranial pressure (ICP) of 117 TBI patients were utilized as the internal validation dataset. The SIH event was explored through the prognostic power of the IH event variables for the outcome after 6 months, and an IH event with thresholds that included an ICP of 20 mmHg and PTD > 130 mmHg * minutes was considered an SIH event. The physiological characteristics of normal, IH and SIH events were investigated. LightGBM was employed to forecast an SIH event from various time intervals using physiological parameters derived …

OBJECTIVE Gray matter (GM) and white matter (WM) are vulnerable to ischemic-edematous insults after traumatic brain injury (TBI). The extent of secondary insult after brain injury is quantifiable using quantitative CT analysis. One conventional quantitative CT measure, the gray-white matter ratio (GWR), and a more recently proposed densitometric analysis are used to assess the extent of these insults. However, the prognostic capacity of the GWR in patients with TBI has not yet been validated. This study aims to test the prognostic value of the GWR and evaluate the alternative parameters derived from the densitometric analysis acquired during the acute phase of TBI. In addition, the prognostic ability of the conventional TBI prognostic models (i.e., IMPACT [International Mission for Prognosis and Analysis of Clinical Trials in TBI] and CRASH [Corticosteroid Randomisation After Significant Head Injury …

Electroencephalogram (EEG)-based applications often require numerous channels to achieve high performance, which limits their widespread use. Various channel selection methods have been proposed to identify minimum EEG channels without compromising performance. However, most methods are limited to specific data paradigms or prediction models. We propose NeuroXAI, a novel method that identifies channel importance regardless of the type of EEG application. It integrates the surrogate analysis algorithm to optimize EEG signals and the data sampling algorithm, which effectively selects from highly voluminous EEG data. The efficacy of channel selection via the proposed method was evaluated through three datasets acquired under different paradigms (motor imagery, steady-state visually evoked potentials, and event-related potentials). On datasets based on these paradigms, NeuroXAI-based …

Sleep stage classification is crucial in diagnosing sleep disorders and monitoring treatment effectiveness, yet it is inconvenient, requiring many electrodes and labor-intensive assessments. Single electroencephalogram (EEG)-based deep learning approaches were proposed to address this issue; however, existing studies have employed deep learning models that are devised for imaging and natural language processing for automated sleep stage classification, which limits their use for capturing sleep patterns or changes over time in EEG signals. This study proposes SleepExpertNet, which is inspired by actual clinical guidelines for scoring sleep stages. The model consists of a representation learning part, which learns the physiological characteristics of EEG sleep signals, and a temporal context modeling part, which learns long- and short-term temporal context information similar to a sleep expert …

Classification of sleep stages is important for diagnosis and treatment of sleep disorder. Manual classification performed by sleep experts is burdensome and time-consuming. This study proposes a novel model for sleep stage classification. EEG and EOG signals of 153 healthy subjects was used. The proposed model ensembles two EEGNet-BiLSTM models which learn EEG and EOG respectively. Compared to the existing models, the two models yielded approximately 82% accuracy and 0.78 k-value, whereas the proposed ensemble model showed 90% accuracy and 0.80 k-value. The proposed ensemble model is superior in terms of accuracy and consistency compared to the conventional models.

Finite element analysis (FEA) is increasingly used to investigate the brain under various pathological changes. Although FEA has been used to study hydrocephalus for decades, previous studies have primarily focused on ventriculomegaly. The present study aimed to investigate the pathologic changes regarding sulcal deformation in normal pressure hydrocephalus (NPH). Two finite element (FE) models—an anatomical brain geometric (ABG) model and the conventional simplified brain geometric (SBG) model—of NPH were constructed. The models were constructed with identical boundary conditions but with different geometries. The ABG model contained details of the sulci geometry, whereas these details were omitted from the SBG model. The resulting pathologic changes were assessed via four biomechanical parameters: pore pressure, von Mises stress, pressure, and void ratio. NPH was induced by …

Background The dynamic relationship between pulse waveform of intracranial pressure (ICP) and transcranial Doppler (TCD) cerebral blood flow velocity (CBFV) may contain information about cerebrospinal compliance. This study investigated the possibility by focusing on the phase shift between fundamental harmonics of CBFV and ICP. Methods Thirty-seven normal pressure hydrocephalus patients (20 men, mean age 58) underwent the cerebrospinal fluid (CSF) infusion tests. The infusion was performed via pre-implanted Ommaya reservoir. The TCD FV was recorded in the middle cerebral artery. Resulting continuous ICP and pressure-volume (PV) signals were analyzed by ICM+ software. Results In initial stage of the CSF infusion, the phase shift was negative (median value = −11°, range = +60 to −117). There was significant inverse association of phase shift with brain elasticity (R = −0.51; p = 0.0009 …

Brain Lab

교수님 소개

김동주

연구실 소개

연구실 키워드

알럼나이가 간 회사

연구실 지원 방법

대우 조건

연도별 논문 인용수

저널 & 학회 논문 출판 수

연구실 논문 목록

NeuroXAI: Adaptive, robust, explainable surrogate framework for determination of channel importance in EEG application

NeuroXAI: Adaptive, robust, explainable surrogate framework for determination of channel importance in EEG application

Hyperosmolar therapy response in traumatic brain injury: Explainable artificial intelligence based long-term time series forecasting approach

연구실 보유 특허

대학별 입시끈

우리 연구실 모집

교수님 소개

김동주

연구실 키워드

알럼나이가 간 회사

연구실 지원 방법

대우 조건

연도별 논문 인용수

저널 & 학회 논문 출판 수

연구실 논문 목록

NeuroXAI: Adaptive, robust, explainable surrogate framework for determination of channel importance in EEG application

NeuroXAI: Adaptive, robust, explainable surrogate framework for determination of channel importance in EEG application

Hyperosmolar therapy response in traumatic brain injury: Explainable artificial intelligence based long-term time series forecasting approach

연구실 보유 특허

최근 3년간 연구실적 높은 연구실

대학별 입시끈

우리 연구실 모집