In our pilot study, we exposed third-trimester fetuses, from week 34 of gestation onwards, twice daily to a maternal spoken nursery rhyme. Two and five weeks after birth, 34 newborns, who were either familiarized with rhyme stimulation in utero or stimulation naïve, were (re-)exposed to the familiar, as well as to a novel and unfamiliar, rhyme, both spoken with the maternal and an unfamiliar female voice. For the stimulation-naïve group, both rhymes were unfamiliar. During stimulus presentation, heart rate activity and high-density electroencephalography were collected and newborns’ responses during familiar and unfamiliar stimulation were analyzed. All newborns demonstrated stronger speech–brain coupling at 1 Hz during the presentation of the maternal voice vs. the unfamiliar female voice. Rhyme familiarity originating from prenatal exposure had no effect on speech–brain coupling in experimentally stimulated newborns. Furthermore, only stimulation-naïve newborns demonstrated an increase in heart rate during the presentation of the unfamiliar female voice. The results indicate prenatal familiarization to auditory speech and point to the specific significance of the maternal voice already in two- to five-week-old newborns.

Human newborns spend up to 18 hours sleeping. The organization of their sleep differs immensely from adult sleep, and its quick maturation and fundamental changes correspond to the rapid cortical development at this age. Manual sleep classification is specifically challenging in this population given major body movements and frequent shifts between vigilance states; in addition various staging criteria co-exist. In the present study we utilized a machine learning approach and investigated how EEG complexity and sleep stages evolve during the very first weeks of life. We analyzed 42 full-term infants which were recorded twice (at week two and five after birth) with full polysomnography. For sleep classification EEG signal complexity was estimated using multi-scale permutation entropy and fed into a machine learning classifier. Interestingly the baby's brain signal complexity (and spectral power) revealed developmental changes in sleep in the first 5 weeks of life, and were restricted to NREM ("quiet") and REM ("active sleep") states with little to no changes in state wake. Data demonstrate that our classifier performs well over chance (i.e., >33% for 3-class classification) and reaches almost human scoring accuracy (60% at week-2, 73% at week-5). Altogether, these results demonstrate that characteristics of newborn sleep develop rapidly in the first weeks of life and can be efficiently identified by means of machine learning techniques.

Sleep has been proposed to indicate preserved residual brain functioning in patients suffering from disorders of consciousness (DOC) after awakening from coma. However, a reliable characterization of sleep patterns in this clinical population continues to be challenging given severely altered brain oscillations, frequent and extended artifacts in clinical recordings and the absence of established staging criteria. In the present study, we try to address these issues and investigate the usefulness of a multivariate machine learning technique based on permutation entropy, a complexity measure. Specifically, we used long-term polysomnography (PSG), along with video recordings in day and night periods in a sample of 23 DOC; 12 patients were diagnosed as Unresponsive Wakefulness Syndrome (UWS) and 11 were diagnosed as Minimally Conscious State (MCS). Eight hour PSG recordings of healthy sleepers (N = 26) were additionally used for training and setting parameters of supervised and unsupervised model, respectively. In DOC, the supervised classification (wake, N1, N2, N3 or REM) was validated using simultaneous videos which identified periods with prolonged eye opening or eye closure. The supervised classification revealed that out of the 23 subjects, 11 patients (5 MCS and 6 UWS) yielded highly accurate classification with an average F1-score of 0.87 representing high overlap between the classifier predicting sleep (i.e. one of the 4 sleep stages) and closed eyes. Furthermore, the unsupervised approach revealed a more complex pattern of sleep-wake stages during the night period in the MCS group, as evidenced by the presence of several distinct clusters. In contrast, in UWS patients no such clustering was found. Altogether, we present a novel data-driven method, based on machine learning that can be used to gain new and unambiguous insights into sleep organization and residual brain functioning of patients with DOC.

You have accessJournal of UrologyStone Disease: New Technology1 Apr 2015PD42-03 ROBUST AUTOMATIC RENAL STONE DETECTION IN ULTRASONIC LIVE STREAMS FOR IMPROVING EXTRACORPOREAL SHOCK WAVE THERAPY Werner Pomwenger, Peter Ott, Stefan Wegenkittl, Reinhold Zimmermann, Olaf Gleibe, and Axel Koch Werner PomwengerWerner Pomwenger More articles by this author , Peter OttPeter Ott More articles by this author , Stefan WegenkittlStefan Wegenkittl More articles by this author , Reinhold ZimmermannReinhold Zimmermann More articles by this author , Olaf GleibeOlaf Gleibe More articles by this author , and Axel KochAxel Koch More articles by this author View All Author Informationhttps://doi.org/10.1016/j.juro.2015.02.2588AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookTwitterLinked InEmail INTRODUCTION AND OBJECTIVES As the gold standard, Extracorporeal Shock Wave Lithotripsy (ESWL) predominantly uses X-ray as a localization system. Ultrasound (US) for stone detection is an alternative method allowing for continuous assessment during the treatment while avoiding radiation exposure. Despite its evident advantages, comparably lower image quality and staff training costs have been identified as major barriers against a wider acceptance. We aim to develop a reliable automatic US-based renal stone detection system to support urologists in ESWL treatment by accurately locating targets in real-time. METHODS Our prototype combines i) a single, frame-based multistage candidate recognition system with ii) a multiple frame-based probabilistic tracking. For i), image enhancement is followed by feature extraction to identify stone candidates. Subsequently, these are examined for the presence of shadowing effects and classified using a finite state machine. The high detection rate (52% to 78%) is achieved by ii) posterior probabilities based on accumulated stone evidence in successive US frames in order to substantially increase sensitivity. To demonstrate robustness, the prototype is evaluated against a set of 5 recorded US sequences (Aloka 3500SX, 3.5MHz transducer, average 60 sec, 17 fps ), annotated by a medical expert, well-trained in US-based stone diagnosis. RESULTS The recorded video sequences showed a strongly varying image quality and typical artefacts (e.g. speckle, ring down, rib shadow or electronic spiking), which was intended as it reflects the real case of clinical application. The correct detection rates vary from 61.7 – 95.0% with sensitivity being from 96.4 – 98.1% with respect to the established ground truth. In this context the positive predictive value was 86.8% in average. CONCLUSIONS As future medical guidelines are expected to call for higher avoidance of radiation exposure of patients, the use of US in ESWL will become desirable. To bring US on a par with X-ray based location, it will strongly benefit from an automatic stone detection system. By combining single- and multi-frame features of US and analyzing these in a decision logic, a substantial increase in stone detection rates was achieved. The moderate to low US video quality used in the tests raises our expectations with respect to the currently ongoing clinical evaluation. With this real-time system US-only based ESWL will perform better in regards of radiation hygiene, stone identification, focus accuracy and treatment efficiency. © 2015 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 193Issue 4SApril 2015Page: e886 Advertisement Copyright & Permissions© 2015 by American Urological Association Education and Research, Inc.MetricsAuthor Information Werner Pomwenger More articles by this author Peter Ott More articles by this author Stefan Wegenkittl More articles by this author Reinhold Zimmermann More articles by this author Olaf Gleibe More articles by this author Axel Koch More articles by this author Expand All Advertisement Advertisement PDF downloadLoading ...