Dynamic functional connectivity in temporal lobe epilepsy: a graph theoretical and machine learning approach

Purpose

Functional magnetic resonance imaging (fMRI) in resting state can be used to evaluate the functional organization of the human brain in the absence of any task or stimulus. The functional connectivity (FC) has non-stationary nature and consented to be varying over time. By considering the dynamic characteristics of the FC and using graph theoretical analysis and a machine learning approach, we aim to identify the laterality in cases of temporal lobe epilepsy (TLE).

Methods

Six global graph measures are extracted from static and dynamic functional connectivity matrices using fMRI data of 35 unilateral TLE subjects. Alterations in the time trend of the graph measures are quantified. The random forest (RF) method is used for the determination of feature importance and selection of dynamic graph features including mean, variance, skewness, kurtosis, and Shannon entropy. The selected features are used in the support vector machine (SVM) classifier to identify the left and right epileptogenic sides in patients with TLE.

Results

Our results for the performance of SVM demonstrate that the utility of dynamic features improves the classification outcome in terms of accuracy (88.5% for dynamic features compared with 82% for static features). Selecting the best dynamic features also elevates the accuracy to 91.5%.

Conclusion

Accounting for the non-stationary characteristics of functional connectivity, dynamic connectivity analysis of graph measures along with machine learning approach can identify the temporal trend of some specific network features. These network features may be used as potential imaging markers in determining the epileptogenic hemisphere in patients with TLE.

     

Is High Preprocedural Renal Resistive Index Sensitive Enough to Predict Iodine Contrast-Induced Nephropathy in Patients Receiving Intra-Arterial Iodinate Contrast?

PurposeRenal Resistive Index (RRI) is a newly introduced sonographic index in predicting contrast-induced nephropathy (CIN) development. It has been suggested that RRI > 0.69 should be considered as a risk factor for CIN development. The present study aimed to calculate the predictive value of RRI using a cutoff point of 0.69.MethodsA total of 90 patients who were a candidate for coronary vessels angiography were enrolled in this study. Color Doppler ultrasonography was performed and RRI was measured. Patients were followed up for 48 hours after contrast media exposure for the CIN development. The diagnosis of CIN was based on a 25% relative rise or 0.5 mg/dL absolute rise in creatinine level. The predictive values of RRI were measured using 0.69 as a cutoff point.ResultsOut of 90 patients, CIN developed in 3 patients and 17 patients had preprocedural RRI > 0.69. Of 3 patients with CIN, 1 had RRI > 0.69. Using 0.69 as the cutoff point, the measured sensitivity and specificity of RRI were 33.3% and 83.9%, respectively.ConclusionsRRI > 0.69 is not a sensitive index in predicting the CIN development and cannot be used as an independent factor.     

Meta-analysis of telomere length in 19?713 subjects reveals high heritability,stronger maternal inheritance and a paternal age effect

Telomere length (TL) has been associated with aging and mortality, but individual differences are also influenced by genetic factors, with previous studies reporting heritability estimates ranging from 34 to 82%. Here we investigate the heritability, mode of inheritance and the influence of parental age at birth on TL in six large, independent cohort studies with a total of 19 713 participants. The meta-analysis estimate of TL heritability was 0.70 (95% CI 0.64–0.76) and is based on a pattern of results that is highly similar for twins and other family members. We observed a stronger mother–offspring (r=0.42; P-value=3.60 × 10⁻⁶¹) than father–offspring correlation (r=0.33; P-value=7.01 × 10⁻⁵), and a significant positive association with paternal age at offspring birth (β=0.005; P-value=7.01 × 10⁻⁵). Interestingly, a significant and quite substantial correlation in TL between spouses (r=0.25; P-value=2.82 × 10⁻³⁰) was seen, which appeared stronger in older spouse pairs (mean age ≥55 years; r=0.31; P-value=4.27 × 10⁻²³) than in younger pairs (mean age<55 years; r=0.20; P-value=3.24 × 10⁻¹⁰). In summary, we find a high and very consistent heritability estimate for TL, evidence for a maternal inheritance component and a positive association with paternal age.     

A novel non-linear recursive filter design for extracting high rate pulse features in nuclear medicine imaging and spectroscopy

Applications in imaging and spectroscopy rely on pulse processing methods for appropriate data generation. Often, the particular method utilized does not highly impact data quality, whereas in some scenarios, such as in the presence of high count rates or high frequency pulses, this issue merits extra consideration. In the present study, a new approach for pulse processing in nuclear medicine imaging and spectroscopy is introduced and evaluated. The new non-linear recursive filter (NLRF) performs nonlinear processing of the input signal and extracts the main pulse characteristics, having the powerful ability to recover pulses that would ordinarily result in pulse pile-up. The filter design defines sampling frequencies lower than the Nyquist frequency.In the literature, for systems involving NaI(Tl) detectors and photomultiplier tubes (PMTs), with a signal bandwidth considered as 15 MHz, the sampling frequency should be at least 30 MHz (the Nyquist rate), whereas in the present work, a sampling rate of 3.3 MHz was shown to yield very promising results. This was obtained by exploiting the known shape feature instead of utilizing a general sampling algorithm. The simulation and experimental results show that the proposed filter enhances count rates in spectroscopy. With this filter, the system behaves almost identically as a general pulse detection system with a dead time considerably reduced to the new sampling time (300 ns). Furthermore, because of its unique feature for determining exact event times, the method could prove very useful in time-of-flight PET imaging.     

Effect of age and sex on fully automated deep learning assessment of left ventricular function,volumes, and contours in cardiac magnetic resonance imaging

The International Journal of Cardiovascular Imaging - Deep learning algorithms for left ventricle (LV) segmentation are prone to bias towards the training dataset. This study assesses sex- and...     

Fixed‐structure controller design for polytopic systems via LMIs

In this paper, a new approach for fixed‐structure H₂ controller design in terms of solutions to a set of linear matrix inequalities are given. Both discrete‐time and continuous‐time SISO time‐invariant systems are considered. Then the results are extended to systems with polytopic uncertainty. The presented methods are based on an inner convex approximation of the non‐convex set of fixed‐structure H₂ controllers. The designed procedures are initialized either with a stable polynomial or with a stabilizing controller. An iterative procedure for robust controller design is given that converges to a suboptimal solution. The monotonic decreasing of the upper bound on the H₂ norm is established theoretically for both nominal and robust controller design. Copyright © 2014 John Wiley & Sons, Ltd.