A multi-label approach using binary relevance and decision trees applied to functional genomics

Many classification problems, especially in the field of bioinformatics, are associated with more than one class, known as multi-label classification problems. In this study, we propose a new adaptation for the Binary Relevance algorithm taking into account possible relations among labels, focusing on the interpretability of the model, not only on its performance. Experiments were conducted to compare the performance of our approach against others commonly found in the literature and applied to functional genomic datasets. The experimental results show that our proposal has a performance comparable to that of other methods and that, at the same time, it provides an interpretable model from the multi-label problem.     

Differentiating ictal/subclinical spikes and waves in childhood absence epilepsy by spectral and network analyses: A pilot study

ObjectiveChildhood absence epilepsy (CAE) is a disease with distinct seizure semiology and electroencephalographic (EEG) features. Differentiating ictal and subclinical generalized spikes and waves discharges (GSWDs) in the EEG is challenging, since they appear to be identical upon visual inspection. Here, spectral and functional connectivity (FC) analyses were applied to routine EEG data of CAE patients, to differentiate ictal and subclinical GSWDs.MethodsTwelve CAE patients with both ictal and subclinical GSWDs were retrospectively selected for this study. The selected EEG epochs were subjected to frequency analysis in the range of 1–30 Hz. Further, FC analysis based on the imaginary part of coherency was used to determine sensor level networks.ResultsDelta, alpha and beta band frequencies during ictal GSWDs showed significantly higher power compared to subclinical GSWDs. FC showed significant network differences for all frequency bands, demonstrating weaker connectivity between channels during ictal GSWDs.ConclusionUsing spectral and FC analyses significant differences between ictal and subclinical GSWDs in CAE patients were detected, suggesting that these features could be used for machine learning classification purposes to improve EEG monitoring.SignificanceIdentifying differences between ictal and subclinical GSWDs using routine EEG, may improve understanding of this syndrome and the management of patients with CAE.     

杜长海名老中医应用药对治疗功能性便秘的临床经验探析

功能性便秘的发生严重困扰着现代人群,进而影响人们的生活质量。杜长海名老中医认为功能性便秘究其内因为大肠传导失常,外因乃现代人的生活方式、紧张的生活氛围等等,因此杜老结合自己数十年的临床经验,根据中医辨证总结出临床常用药对,对有效治疗功能性便秘进行了经验总结。     

Is watch and wait a safe and effective way to treat rectal cancer in older patients?

IntroductionThe aim was assess the oncological and functional outcome of the watch-and-wait (W&W) approach in older patients with a clinical (near)complete response after neoadjuvant treatment for rectal cancer.Material and methodsPatients were included in a W&W-approach (2004–2019) when digital rectal examination, endoscopy and MRI showed a (near)clinical complete response. Patients underwent endoscopy and MRI every 3 months during the first year, and 6-monthly thereafter. Patients aged ≥75 and ≥ 2 years of follow-up (FU) were selected. Oncological outcomes were assessed with Kaplan-Meier curves. Functional outcome was assessed with colostomy-free rate, Vaizey incontinence score, low anterior resection syndrome-score and International Prostate Syndrome Score.Results43/304 (14%) of patients in a W&W-approach met the inclusion criteria. Median FU was 37 (24–109) months. 5/43(12%) developed a local regrowth. All were treated surgically, with one patient experiencing a pelvic failure. Distant metastases occurred in 3/43 patients and 4 patients died, 3 of whom not related to rectal cancer. The 3-year local regrowth-free rate was 88%, 3-year non-regrowth disease-free survival 91%, overall survival 97% and 3-year colostomy-free rate 93%. Overall, the bowel- and urinary dysfunction scores at 3, 12 and 24 months indicated good continence, no or minor LARS and moderate urinary problems.ConclusionW&W for older patients with a clinical (near) complete response appears to be a safe alternative to a total mesorectal excision (TME), with a very high pelvic control rate, and few rectal cancer related deaths. Most patients can avoid major surgery and a definitive colostomy, and have a reasonable anorectal and urinary function.     

Basal ganglia and cerebellar circuits have distinct roles in blepharospasm

IntroductionTo identify areas of brain activity associated with involuntary muscle contractions in patients with blepharospasm using functional MRI.Methods15 patients with blepharospasm underwent 8-min resting state scans with spontaneous orbicularis oculi muscle contractions simultaneously recorded using MRI-compatible surface electromyography. Spasm severity and spasm onset/offset were modeled using the amplitude of the electromyography signal (EMG-Amp) and its first temporal derivative (EMG-Onset), respectively, and included in a multiple regression functional MRI analysis using SPM12. Primary outcome was within-group blood-oxygen-level dependent activations that co-varied with EMG-Amp and EMG-Onset following correction for multiple comparisons for an overall cluster corrected p < 0.05. Secondary analyses included testing for correlations between imaging findings and symptom severity, as measured by clinical dystonia rating scales, using an uncorrected voxel-level threshold of p < 0.001.ResultsImaging data from one subject were excluded due to excessive movement. EMG-Amp co-activated within the left sensorimotor cortex and cerebellum, as well as right lingual gyrus and superior temporal gyrus. EMG-Onset co-activated within the left posterior putamen/pallidum and a frontal eye field region in the left superior frontal gyrus. Symptom severity and EMG-Amp significantly co-varied in a small cluster within the left cerebellum.ConclusionOur preliminary findings here suggest that cerebello-cortical circuits in blepharospasm could drive the intensity of eyelid spasms while basal ganglia circuits are associated with the triggering of spasms. This supports the network model for dystonia and identifies specific areas of involvement consistent with known brain regions responsible for control of movement.