Connecting cohorts of Finnish biobanks creates a research resource for the study of healthy ageing.

AIMS: Connecting cohorts with biobanks is a Finnish biobank collaboration, creating an infrastructure for the study of healthy ageing. We aimed to develop a model for data integration and harmonisation between different biobanks with procedures for joint access. METHODS: The heart of the collaboration is the integrated datasets formed by using data from three biobanks: (a) Arctic Biobank, hosting regional birth cohorts and cohorts of elderly; (b) hospital-affiliated Borealis Biobank of Northern Finland; and (c) THL Biobank, hosting population-based cohorts. The datasets were created by developing a data dictionary, harmonising cohort data and with a joint pseudonymisation process. RESULTS: The connecting cohorts with biobanks resource at its widest consists altogether of almost 1.4 million individuals from collaborating biobanks. Utilising data from 107,000 cohort participants, we created harmonised datasets that contain attributes describing metabolic risk and frailty for studies of healthy ageing. These data can be complemented with medical data available from Biobank Borealis and with samples taken at hospital settings for approximately 38,000 cohort participants. In addition, the harmonised connecting cohorts with biobanks datasets can be expanded with supplementary data and samples from the collaborating biobanks. CONCLUSIONS: The connecting cohorts with biobanks datasets provide a unique resource for research on ageing-related personalised healthcare and for real-world evidence studies. Following the FAIR principles on findability, accessibility, interoperability, and reusability, the reused and harmonised datasets are findable and made accessible for researchers. The same approach can be further utilised to develop additional datasets for other research topics.

Prognostic value of P-wave morphology in general population.

AIMS: To evaluate the prognostic significance of novel P-wave morphology descriptors in general population. METHODS AND RESULTS: Novel P-wave morphology variables were analyzed from orthogonal X-, Y-, Z-leads of the digitized electrocardiogram using a custom-made software in 6906 middle-aged subjects of the Mini-Finland Health Survey. A total of 3747 (54.3%) participants died during the follow-up period of 24.3 ± 10.4 years; 379 (5.5%) of the study population succumbed to sudden cardiac death (SCD), 928 (13.4%) to non-SCD (NSCD) and 2440 (35.3%) patients to non-cardiac death (NCD). In univariate comparisons, most of the studied P-wave morphology parameters had a significant association with all modes of death (P from <0.05 to <0.001). After relevant adjustments in the Cox multivariate hazards model, P-wave morphology dispersion (PMD) still tended to predict SCD [hazard ratio (HR): 1.006, 95% confidence interval (CI): 1.000-1.012, P = 0.05) but not NSCD (HR: 0.999, 95% CI: 0.995-1.003, P = 0.68) or NCD (HR: 0.999, 95% CI: 0.997-1.001, P = 0.44). The P-wave maximum amplitude in the lead Z (P-MaxAmp-Z) predicted SCD even after multivariate adjustments (HR: 1.010, 95% CI: 1.005-1.015, P = 0.0002) but also NSCD (HR: 1.005, 95% CI: 1.002-1.009, P = 0.0005) and NCD (HR: 1.002, 95% CI: 1.000-1.005, P = 0.03). CONCLUSION: Abnormalities of P-wave morphology are associated with the risk of all modes of death in general population. After relevant adjustments, PMD was still closely associated with the risk of SCD but not with NSCD or NCD. P-MaxAmp-Z predicted SCD even after adjustments, however, it also retained its association with NSCD and NCD.

Cardiovascular brain impulses in Alzheimer's disease.

Accumulation of amyloid-ß is a key neuropathological feature in brain of Alzheimer's disease patients. Alterations in cerebral haemodynamics, such as arterial impulse propagation driving the (peri)vascular CSF flux, predict future Alzheimer's disease progression. We now present a non-invasive method to quantify the three-dimensional propagation of cardiovascular impulses in human brain using ultrafast 10 Hz magnetic resonance encephalography. This technique revealed spatio-temporal abnormalities in impulse propagation in Alzheimer's disease. The arrival latency and propagation speed both differed in patients with Alzheimer's disease. Our mapping of arterial territories revealed Alzheimer's disease-specific modifications, including reversed impulse propagation around the hippocampi and in parietal cortical areas. The findings imply that pervasive abnormality in (peri)vascular CSF impulse propagation compromises vascular impulse propagation and subsequently glymphatic brain clearance of amyloid-ß in Alzheimer's disease.

SAM: Self-augmentation mechanism for COVID-19 detection using chest X-ray images.

COVID-19 is a rapidly spreading viral disease and has affected over 100 countries worldwide. The numbers of casualties and cases of infection have escalated particularly in countries with weakened healthcare systems. Recently, reverse transcription-polymerase chain reaction (RT-PCR) is the test of choice for diagnosing COVID-19. However, current evidence suggests that COVID-19 infected patients are mostly stimulated from a lung infection after coming in contact with this virus. Therefore, chest X-ray (i.e., radiography) and chest CT can be a surrogate in some countries where PCR is not readily available. This has forced the scientific community to detect COVID-19 infection from X-ray images and recently proposed machine learning methods offer great promise for fast and accurate detection. Deep learning with convolutional neural networks (CNNs) has been successfully applied to radiological imaging for improving the accuracy of diagnosis. However, the performance remains limited due to the lack of representative X-ray images available in public benchmark datasets. To alleviate this issue, we propose a self-augmentation mechanism for data augmentation in the feature space rather than in the data space using reconstruction independent component analysis (RICA). Specifically, a unified architecture is proposed which contains a deep convolutional neural network (CNN), a feature augmentation mechanism, and a bidirectional LSTM (BiLSTM). The CNN provides the high-level features extracted at the pooling layer where the augmentation mechanism chooses the most relevant features and generates low-dimensional augmented features. Finally, BiLSTM is used to classify the processed sequential information. We conducted experiments on three publicly available databases to show that the proposed approach achieves the state-of-the-art results with accuracy of 97%, 84% and 98%. Explainability analysis has been carried out using feature visualization through PCA projection and t-SNE plots.

Prognostic significance of flat T-waves in the lateral leads in general population.

BACKGROUND: Negative T-waves are associated with sudden cardiac death (SCD) risk in the general population. Whether flat T-waves also predict SCD is not known. The aim of the study was to examine the clinical characteristics and risk of SCD in general population subjects with flat T-waves. METHODS: We examined the electrocardiograms of 6750 Finnish general population adults aged ≥30 years and classified the subjects into 3 groups: 1) negative T-waves with an amplitude ≥0.1 mV in ≥2 of the leads I, II, aVL, V4-V6, 2) negative or positive low amplitude T-waves with an amplitude <0.1 mV and the ratio of T-wave and R-wave <10% in ≥2 of the leads I, II, aVL, V4-V6, and 3) normal positive T-waves (not meeting the aforesaid criteria). The association between T-wave classification and SCD was assessed during a 10-year follow-up. RESULTS: A total of 215 (3.2%) subjects had negative T-waves, 856 (12.7%) flat T-waves, and 5679 (84.1%) normal T-waves. Flat T-wave subjects were older and had more often cardiovascular morbidities compared to normal T-wave subjects, while negative T-wave subjects were the oldest and had most often cardiovascular morbidities. After adjusting for multiple factors, both flat T-waves (hazard ratio [HR] 1.81; 95% confidence interval [CI] 1.13-2.91) and negative T-waves (HR 3.27; 95% CI 1.85-5.78) associated with SCD. CONCLUSIONS: Cardiovascular risk factors and disease are common among subjects with flat T-waves, but these minor T-wave abnormalities are also independently associated with increased SCD risk.

Joint Image Encryption and Screen-Cam Robust Two Watermarking Scheme.

This paper proposes a joint encryption and screen-cam robust watermarking scheme. This method combines the advantages of smartphone, encryption and watermarking technologies, thereby achieving watermark extraction with a smartphone, partial decryption and tracking leakage from sneak shots. We design a dual watermarking algorithm to achieve watermark detection from both encrypted and decrypted images. First, a watermark is embedded in the discrete Fourier transform (DFT) domain to enable leakage tracking. Then, a second watermark is generated based on QR (Quick response) code encoding and inverse DFT to achieve high watermark capacity and error correction ability, where the secret key for decryption is included in the watermark message. By hiding this message carrying the watermark for the encrypted image in the changes caused by embedding the first watermark, we can improve imperceptibility and will not affect the effectiveness of the proposed scheme. Finally, to enhance the robustness of watermark after encryption, a chaotic mapping-based segment encryption algorithm is proposed. In the process of watermark detection, to cope with perspective correction, a frame locating based algorithm is employed to achieve watermark synchronization from a recaptured picture of the encrypted image. Considering the severe quality degradation, we use a noise component and local statistic feature-based method to extract the message bits. The experimental results show that the proposed scheme is secure, and highly robust, to screen-cam the process for both before and after decryption. Additionally, after decryption, the proposed scheme also has high robustness against common image processing attacks.

Orthogonal P-wave morphology, conventional P-wave indices, and the risk of atrial fibrillation in the general population using data from the Finnish Hospital Discharge Register.

AIMS: Identifying subjects at high and low risk of atrial fibrillation (AF) is of interest. This study aims to assess the risk of AF associated with electrocardiographic (ECG) markers linked to atrial fibrosis: P-wave prolongation, 3rd-degree interatrial block, P-terminal force in lead V1, and orthogonal P-wave morphology. METHODS AND RESULTS: P-wave parameters were assessed in a representative Finnish population sample aged ≥30 years (n = 7217, 46.0% male, mean age 51.4 years). Subjects (n = 5489) with a readable ECG including the orthogonal leads, sinus rhythm, and a predefined orthogonal P-wave morphology type [positive in leads X and Y and either negative (Type 1) or ± biphasic (Type 2) in lead Z; Type 3 defined as positive in lead X and ± biphasic in lead Y], were followed 10 years from the baseline examinations (performed 1978-80). Subjects discharged with AF diagnosis after any-cause hospitalization (n = 124) were defined as having developed AF. Third-degree interatrial block was defined as P-wave ≥120 ms and the presence of ≥2 ± biphasic P waves in the inferior leads. Hazard ratios (HRs) and confidence intervals (CIs) were assessed with Cox models. Third-degree interatrial block (n = 103, HR 3.18, 95% CI 1.66-6.13; P = 0.001) and Type 3 morphology (n = 216, HR 3.01, 95% CI 1.66-5.45; P < 0.001) were independently associated with the risk of hospitalization with AF. Subjects with P-wave <110 ms and Type 1 morphology (n = 2074) were at low risk (HR 0.46, 95% CI 0.26-0.83; P = 0.006), compared to the rest of the subjects. CONCLUSION: P-wave parameters associate with the risk of hospitalization with AF.

A Video-Based DT-SVM School Violence Detecting Algorithm.

School bullying is a serious problem among teenagers. School violence is one type of school bullying and considered to be the most harmful. As AI (Artificial Intelligence) techniques develop, there are now new methods to detect school violence. This paper proposes a video-based school violence detecting algorithm. This algorithm first detects foreground moving targets via the KNN (K-Nearest Neighbor) method and then preprocesses the detected targets via morphological processing methods. Then, this paper proposes a circumscribed rectangular frame integrating method to optimize the circumscribed rectangular frame of moving targets. Rectangular frame features and optical-flow features were extracted to describe the differences between school violence and daily-life activities. We used the Relief-F and Wrapper algorithms to reduce the feature dimension. SVM (Support Vector Machine) was applied as the classifier, and 5-fold cross validation was performed. The accuracy was 89.6%, and the precision was 94.4%. To further improve the recognition performance, we developed a DT-SVM (Decision Tree-SVM) two-layer classifier. We used boxplots to determine some features of the DT layer that are able to distinguish between typical physical violence and daily-life activities and between typical daily-life activities and physical violence. For the remainder of activities, the SVM layer performed a classification. For this DT-SVM classifier, the accuracy reached 97.6%, and the precision reached 97.2%, thus showing a significant improvement.

Musculoskeletal pains and cardiovascular autonomic function in the general Northern Finnish population.

BACKGROUND: Heart rate variability (HRV) and baroreflex sensitivity (BRS) measurements provide means for the objective assessment of cardiovascular autonomic function. As previous studies have associated chronic pain with abnormal autonomic function, we aimed to characterize the relationship between the number of musculoskeletal pain sites (NPS), pain intensity, and cardiovascular autonomic function among the population-based Northern Finland Birth Cohort 1966. METHODS: At the age of 46, cohort members self-reported their musculoskeletal pains (enabling the determination of NPS [0-8] and pain intensity [Numerical Rating Scale, NRS, 0-10]) and underwent clinical assessments of cardiovascular autonomic function in seated and standing positions (HRV variables: heart rate [HR] and root mean square of successive differences in beat-to-beat intervals [rMSSD] for the entire cohort; BRS variables: low-frequency systolic blood pressure variability [SBPV] and cross-spectral baroreflex sensitivity [BRS] for those attending the examination in Oulu, Finland). Extensive confounder data were also collected (body mass index, physical activity, smoking, Hopkins Symptom Checklist-25, comorbidities, and medications). The full samples included 4186 and 2031 individuals (HRV and BRS samples, respectively). Three subanalyses focused on individuals with intense and frequent pain, individuals with symptoms of depression and anxiety, and the relationship between pain intensity and autonomic parameters. RESULTS: Linear regression models showed varying associations between NPS, pain intensity, and cardiovascular autonomic parameters. However, after all adjustments NPS was only associated with one outcome among women (BRS, standing: beta = - 0.015, p = 0.048) and two among men (HR, seated: beta = - 0.902, p = 0.003; HR, standing: beta = - 0.843, p = 0.014). Pain intensity was not associated with any outcome after full adjustments. Significant sex*pain interactions were found in the data. CONCLUSIONS: Our data suggest that musculoskeletal pain has, at most, a limited independent association with cardiovascular autonomic function. Future studies should carefully account for the potential confounders and sex interactions that this study revealed.

Spectral fusion-based breathing frequency estimation; experiment on activities of daily living.

BACKGROUND: We study the estimation of breathing frequency (BF) derived from wearable single-channel ECG signal in the context of mobile daily life activities. Although respiration effects on heart rate variability and ECG morphology have been well established, studies on ECG-derived respiration in daily living settings are scarce; possibly due to considerable amount of disturbances in such data. Yet, unobtrusive BF estimation during everyday activities can provide vital information for both disease management and athletic performance optimization. METHOD AND DATA: For robust ECG-derived BF estimation, we combine the respiratory information derived from R-R interval (RRI) variability and morphological scale variation of QRS complexes (MSV), acquired from ECG signals. Two different fusion techniques are applied on MSV and RRI signals: cross-power spectral density (CPSD) estimation and power spectrum multiplication (PSM). The algorithms were tested on large sets of data collected from 67 participants during office, household and sport activities, simulating daily living activities. We use spirometer reference BF to evaluate and compare our estimations made by different models. RESULTS AND CONCLUSION: PSM acquires the least average error of BF estimation, [Formula: see text] and [Formula: see text], compared to the reference spirometer values. PSM offers approximately 25 and 75% less error in comparison with the CPSD fusion estimation and the estimation by those two exclusive sources, respectively. Our results demonstrate the superiority of both of the fusion approaches, compared to the estimation derived from either of RRI or MSV signals exclusively.

Experiences in digitizing and digitally measuring a paper-based ECG archive.

BACKGROUND: No established method for digitizing and digital measuring of paper electrocardiograms (ECG) exists. We describe a paper ECG digitizing and digital measuring process, and report comparability to manual measurements. METHODS: A paper ECG was recorded from 7203 health survey participants in 1978-1980. With specific software, the ECGs were digitized (ECG Trace Tool), and measured digitally (EASE). A sub-sample of 100 ECGs was selected for manual measurements. RESULTS: The measurement methods showed good agreement. The mean global (EASE)-(manual) differences were 1.4ms (95% CI 0.5-2.2) for PR interval, -1.0ms (95% CI -1.5-[-0.5]) for QRS duration, and 11.6ms (95% CI 10.5-12.7) for QT interval. The mean inter-method amplitude differences of RampV5, RampV6, SampV1, TampII and TampV5 ranged from -0.03mV to 0.01mV. CONCLUSIONS: The presented paper-to-digital conversion and digital measurement process is an accurate and reliable method, enabling efficient storing and analysis of paper ECGs.

Pilot Study of Propofol-induced Slow Waves as a Pharmacologic Test for Brain Dysfunction after Brain Injury.

BACKGROUND: Slow waves (less than 1 Hz) are the most important electroencephalogram signatures of nonrapid eye movement sleep. While considered to have a substantial importance in, for example, providing conditions for single-cell rest and preventing long-term neural damage, a disturbance in this neurophysiologic phenomenon is a potential indicator of brain dysfunction. METHODS: Since, in healthy individuals, slow waves can be induced with anesthetics, the authors tested the possible association between hypoxic brain injury and slow-wave activity in comatose postcardiac arrest patients (n = 10) using controlled propofol exposure. The slow-wave activity was determined by calculating the low-frequency (less than 1 Hz) power of the electroencephalograms recorded approximately 48 h after cardiac arrest. To define the association between the slow waves and the potential brain injury, the patients' neurologic recovery was then followed up for 6 months. RESULTS: In the patients with good neurologic outcome (n = 6), the low-frequency power of electroencephalogram representing the slow-wave activity was found to substantially increase (mean ± SD, 190 ± 83%) due to the administration of propofol. By contrast, the patients with poor neurologic outcome (n = 4) were unable to generate propofol-induced slow waves. CONCLUSIONS: In this experimental pilot study, the comatose postcardiac arrest patients with poor neurologic outcome were unable to generate normal propofol-induced electroencephalographic slow-wave activity 48 h after cardiac arrest. The finding might offer potential for developing a pharmacologic test for prognostication of brain injury by measuring the electroencephalographic response to propofol.

Real-Time Tissue Classification Using a Novel Optical Needle Probe for Biopsy.

Core needle biopsy is a part of the histopathological process, which is required for cancerous tissue examination. The most common method to guide the needle inside of the body is ultrasound screening, which in greater part is also the only guidance method. Ultrasound screening requires user experience. Furthermore, patient involuntary movements such as breathing might introduce artifacts and blur the screen. Optically enhanced core needle biopsy probe could potentially aid interventional radiologists during this procedure, providing real-time information on tissue properties close to the needle tip, while it is advancing inside of the body. In this study, we used diffuse optical spectroscopy in a custom-made core needle probe for real-time tissue classification. Our aim was to provide initial characteristics of the smart needle probe in the differentiation of tissues and validate the basic purpose of the probe of informing about breaking into a desired organ. We collected optical spectra from rat blood, fat, heart, kidney, liver, lungs, and muscle tissues. Gathered data were analyzed for feature extraction and evaluation of two machine learning-based classifiers: support vector machine and k-nearest neighbors. Their performances on training data were compared using subject-independent k-fold cross-validation. The best classifier model was chosen and its feasibility for real-time automated tissue recognition and classification was then evaluated. The final model reached nearly 80% of correct real-time classification of rat organs when using the needle probe during real-time classification.

Reducing the airflow waveform distortions from breathing style and body position with improved calibration of respiratory effort belts.

BACKGROUND: Respiratory effort belt measurement is a widely used method to monitor respiration. Signal waveforms of respiratory volume and flow may indicate pathological signs of several diseases and, thus, it would be highly desirable to predict them accurately. Calibrated effort belts are sufficiently accurate for estimating respiratory rate, but the respiratory volume and flow prediction accuracies degrade considerably with changes in the subject's body position and breathing style. METHODS: An improved calibration method of respiratory effort belts is presented in this paper. It is based on an optimally trained FIR (Finite Impulse Response) filter bank constructed as a MISO system (Multiple-Input Single-Output) between respiratory effort belt signals and the spirometer in order to reduce waveform errors. Ten healthy adult volunteers were recruited. Breathing was varied between the following styles: metronome-guided controlled breathing rate of 0.1 Hz, 0.15 Hz, 0.25 Hz and 0.33 Hz, and a free rate that was felt normal by each subject. Body position was varied between supine, sitting and standing. The proposed calibration method was tested against these variations and compared with the state-of-the-art methods from the literature. RESULTS: Relative waveform error decreased 60-70% when predicting airflow under changing breathing styles. The coefficient of determination R2 varied between 0.88-0.95 and 0.65-0.79 with the proposed and the standard method, respectively. Standard deviation of respiratory volume error decreased even 80%. The proposed method outperformed other methods. CONCLUSIONS: Results show that not only the respiratory volume can be computed more precisely from the predicted airflow, but also the flow waveforms are very accurate with the proposed method. The method is robust to breathing style changes and body position changes improving greatly the accuracy of the calibration of respiratory effort belts over the standard method. The enhanced accuracy of the belt calibration offers interesting opportunities, e.g. in pulmonary and critical care medicine when objective measurements are required.

Automatic captioning for medical imaging (MIC): a rapid review of literature.

Automatically understanding the content of medical images and delivering accurate descriptions is an emerging field of artificial intelligence that combines skills in both computer vision and natural language processing fields. Medical image captioning is involved in various applications related to diagnosis, treatment, report generation and computer-aided diagnosis to facilitate the decision making and clinical workflows. Unlike generic image captioning, medical image captioning highlights the relationships between image objects and clinical findings, which makes it a very challenging task. Although few review papers have already been published in this field, their coverage is still quite limited and only particular problems are addressed. This motivates the current paper where a rapid review protocol was adopted to review the latest achievements in automatic medical image captioning from the medical domain perspective. We aim through this review to provide the reader with an up-to-date literature in this field by summarizing the key findings and approaches in this field, including the related datasets, applications and limitations as well as highlighting the main competitions, challenges and future directions.

Invasion depth estimation of carcinoma cells using adaptive stain normalization to improve epidermis segmentation accuracy.

Submucosal invasion depth is a significant prognostic factor when assessing lymph node metastasis and cancer itself to plan proper treatment for the patient. Conventionally, oncologists measure the invasion depth by hand which is a laborious, subjective, and time-consuming process. The manual pathological examination by measuring accurate carcinoma cell invasion with considerable inter-observer and intra-observer variations is still challenging. The increasing use of medical imaging and artificial intelligence reveals a significant role in clinical medicine and pathology. In this paper, we propose an approach to study invasive behavior and measure the invasion depth of carcinoma from stained histopathology images. Specifically, our model includes adaptive stain normalization, color decomposition, and morphological reconstruction with adaptive thresholding to separate the epithelium with blue ratio image. Our method splits the image into multiple non-overlapping meaningful segments and successfully finds the homogeneous segments to measure accurate invasion depth. The invasion depths are measured from the inner epithelium edge to outermost pixels of the deepest part of particles in image. We conduct our experiments on skin melanoma tissue samples as well as on organotypic invasion model utilizing myoma tissue and oral squamous cell carcinoma. The performance is experimentally compared to three closely related reference methods and our method provides a superior result in measuring invasion depth. This computational technique will be beneficial for the segmentation of epithelium and other particles for the development of novel computer-aided diagnostic tools in biobank applications.

Optimized detection and segmentation of nuclei in gastric cancer images using stain normalization and blurred artifact removal.

Histological analysis with microscopy is the gold standard to diagnose and stage cancer, where slides or whole slide images are analyzed for cell morphological and spatial features by pathologists. The nuclei of cancerous cells are characterized by nonuniform chromatin distribution, irregular shapes, and varying size. As nucleus area and shape alone carry prognostic value, detection and segmentation of nuclei are among the most important steps in disease grading. However, evaluation of nuclei is a laborious, time-consuming, and subjective process with large variation among pathologists. Recent advances in digital pathology have allowed significant applications in nuclei detection, segmentation, and classification, but automated image analysis is greatly affected by staining factors, scanner variability, and imaging artifacts, requiring robust image preprocessing, normalization, and segmentation methods for clinically satisfactory results. In this paper, we aimed to evaluate and compare the digital image analysis techniques used in clinical pathology and research in the setting of gastric cancer. A literature review was conducted to evaluate potential methods of improving nuclei detection. Digitized images of 35 patients from a retrospective cohort of gastric adenocarcinoma at Oulu University Hospital in 1987-2016 were annotated for nuclei (n = 9085) by expert pathologists and 14 images of different cancer types from public TCGA dataset with annotated nuclei (n = 7000) were used as a comparison to evaluate applicability in other cancer types. The detection and segmentation accuracy with the selected color normalization and stain separation techniques were compared between the methods. The extracted information can be supplemented by patient's medical data and fed to the existing statistical clinical tools or subjected to subsequent AI-assisted classification and prediction models. The performance of each method is evaluated by several metrics against the annotations done by expert pathologists. The F1-measure of 0.854 ± 0.068 is achieved with color normalization for the gastric cancer dataset, and 0.907 ± 0.044 with color deconvolution for the public dataset, showing comparable results to the earlier state-of-the-art works. The developed techniques serve as a basis for further research on application and interpretability of AI-assisted tools for gastric cancer diagnosis.

Respiratory brain impulse propagation in focal epilepsy.

Respiratory brain pulsations pertaining to intra-axial hydrodynamic solute transport are markedly altered in focal epilepsy. We used optical flow analysis of ultra-fast functional magnetic resonance imaging (fMRI) data to investigate the velocity characteristics of respiratory brain impulse propagation in patients with focal epilepsy treated with antiseizure medication (ASM) (medicated patients with focal epilepsy; ME, n = 23), drug-naïve patients with at least one seizure (DN, n = 19) and matched healthy control subjects (HC, n = 75). We detected in the two patient groups (ME and DN) several significant alterations in the respiratory brain pulsation propagation velocity, which showed a bidirectional change dominated by a reduction in speed. Furthermore, the respiratory impulses moved more in reversed or incoherent directions in both patient groups vs. the HC group. The speed reductions and directionality changes occurred in specific phases of the respiratory cycle. In conclusion, irrespective of medication status, both patient groups showed incoherent and slower respiratory brain impulses, which may contribute to epileptic brain pathology by hindering brain hydrodynamics.

Non-Contact Atrial Fibrillation Detection From Face Videos by Learning Systolic Peaks.

OBJECTIVE: We propose a non-contact approach for atrial fibrillation (AF) detection from face videos. METHODS: Face videos, electrocardiography (ECG), and contact photoplethysmography (PPG) from 100 healthy subjects and 100 AF patients are recorded. Data recordings from healthy subjects are all labeled as healthy. Two cardiologists evaluated ECG recordings of patients and labeled each recording as AF, sinus rhythm (SR), or atrial flutter (AFL). We use the 3D convolutional neural network for remote PPG monitoring and propose a novel loss function (Wasserstein distance) to use the timing of systolic peaks from contact PPG as the label for our model training. Then a set of heart rate variability (HRV) features are calculated from the inter-beat intervals, and a support vector machine (SVM) classifier is trained with HRV features. RESULTS: Our proposed method can accurately extract systolic peaks from face videos for AF detection. The proposed method is trained with subject-independent 10-fold cross-validation with 30 s video clips and tested on two tasks. 1) Classification of healthy versus AF: the accuracy, sensitivity, and specificity are 96.00%, 95.36%, and 96.12%. 2) Classification of SR versus AF: the accuracy, sensitivity, and specificity are 95.23%, 98.53%, and 91.12%. In addition, we also demonstrate the feasibility of non-contact AFL detection. CONCLUSION: We achieve good performance of non-contact AF detection by learning systolic peaks. SIGNIFICANCE: non-contact AF detection can be used for self-screening of AF symptoms for suspectable populations at home or self-monitoring of AF recurrence after treatment for chronic patients.

Whole slide image registration via multi-stained feature matching.

In the recent decade, medical image registration and fusion process has emerged as an effective application to follow up diseases and decide the necessary therapies based on the conditions of patient. For many of the considerable diagnostic analyses, it is common practice to assess two or more different histological slides or images from one tissue sample. A specific area analysis of two image modalities requires an overlay of the images to distinguish positions in the sample that are organized at a similar coordinate in both images. In particular cases, there are two common challenges in digital pathology: first, dissimilar appearances of images resulting due to staining variances and artifacts; second, large image size. In this paper, we develop algorithm to overcome the fact that scanners from different manufacturers have variations in the images. We propose whole slide image registration algorithm where adaptive smoothing is employed to smooth the stained image. A modified scale-invariant feature transform is applied to extract common information and a joint distance helps to match keypoints correctly by eliminating position transformation error. Finally, the registered image is obtained by utilizing correct correspondences and the interpolation of color intensities. We validate our proposal using different images acquired from surgical resection samples of lung cancer (adenocarcinoma). Extensive feature matching with apparently increasing correct correspondences and registration performance on several images demonstrate the superiority of our method over state-of-the-art methods. Our method potentially improves the matching accuracy that might be beneficial for computer-aided diagnosis in biobank applications.