It is known that culture media (CM) promotes cellular growth, adhesion, and protects explanted primary brain cells from in vitro stresses. The fetal bovine serum (FBS) supplement used in most CM, however, contains significant quantities of extracellular vesicles (EVs) that confound quantitative and qualitative analyses from the EVs produced by the cultured cells. We quantitatively tested the ability of common FBS EV-depletion protocols to remove exogenous EVs from FBS-supplemented CM and evaluated the influence such methods have on primary astrocyte culture growth and viability. We assessed two methodologies utilized for FBS EV removal prior to adding to CM: (1) an 18-h ultracentrifugation (UC); and (2) a commercial EV-depleted FBS (Exo-FBS™). Our analysis demonstrated that Exo-FBS™ CM provided the largest depletion (75%) of total FBS EVs, while still providing 6.92 × 10⁹ ± 1.39 × 10⁸ EVs/mL. In addition, both UC and Exo-FBS™ CM resulted in poor primary astrocyte cell growth and viability in culture. The two common FBS EV-depletion methods investigated, therefore, not only contaminate in vitro primary cell-derived EV analyses, but also provide a suboptimal environment for primary astrocyte cell growth and viability. It appears likely that future CM optimization, using a serum-free alternative, might be required to advance analyses of cell-specific EVs isolated in vitro.

Image registration is an essential step in many medical image analysis tasks. Traditional methods for image registration are primarily optimization-driven, finding the optimal deformations that maximize the similarity between two images. Recent learning-based methods, trained to directly predict transformations between two images, run much faster, but suffer from performance deficiencies due to domain shift. Here we present a new neural network based image registration framework, called NIR (Neural Image Registration), which is based on optimization but utilizes deep neural networks to model deformations between image pairs. NIR represents the transformation between two images with a continuous function implemented via neural fields, receiving a 3D coordinate as input and outputting the corresponding deformation vector. NIR provides two ways of generating deformation field: directly output a displacement vector field for general deformable registration, or output a velocity vector field and integrate the velocity field to derive the deformation field for diffeomorphic image registration. The optimal registration is discovered by updating the parameters of the neural field via stochastic mini-batch gradient descent. We describe several design choices that facilitate model optimization, including coordinate encoding, sinusoidal activation, coordinate sampling, and intensity sampling. NIR is evaluated on two 3D MR brain scan datasets, demonstrating highly competitive performance in terms of both registration accuracy and regularity. Compared to traditional optimization-based methods, our approach achieves better results in shorter computation times. In addition, our methods exhibit performance on a cross-dataset registration task, compared to the pre-trained learning-based methods.

Objectives: To develop a multiclass-classifier deep learning model and website for distinguishing tympanic membrane (TM) pathologies based on otoscopic images. Methods: An otoscopic image database developed by utilizing publicly available online images and open databases was assessed by convolutional neural network (CNN) models including ResNet-50, Inception-V3, Inception-Resnet-V2, and MobileNetV2. Training and testing were conducted with a 75:25 breakdown. Area under the curve of receiver operating characteristics (AUC-ROC), accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were used to compare different CNN models’ performances in classifying TM images. Results: Our database included 400 images, organized into normal (n = 196) and abnormal classes (n = 204), including acute otitis media (n = 116), otitis externa (n = 44), chronic suppurative otitis media (n = 23), and cerumen impaction (n = 21). For binary classification between normal versus abnormal TM, the best performing model had average AUC-ROC of 0.902 (MobileNetV2), followed by 0.745 (Inception-Resnet-V2), 0.731 (ResNet-50), and 0.636 (Inception-V3). Accuracy ranged between 0.73–0.77, sensitivity 0.72–0.88, specificity 0.58–0.84, PPV 0.68–0.81, and NPV 0.73–0.83. Macro-AUC-ROC for MobileNetV2 based multiclass-classifier was 0.91, with accuracy of 66%. Binary and multiclass-classifier models based on MobileNetV2 were loaded onto a publicly accessible and user-friendly website (https://headneckml.com/tympanic). This allows the readership to upload TM images for real-time predictions using the developed algorithms. Conclusions: Novel CNN algorithms were developed with high AUC-ROCs for differentiating between various TM pathologies. This was further deployed as a proof-of-concept publicly accessible website for real-time predictions.

To develop a smartphone application providing sound therapy and cognitive behavioral therapy (CBT) for treating tinnitus and performing a proof-of-concept pilot study evaluating its potential efficacy. An interactive smartphone application available on iOS and Android platforms was developed, which provided an 8-week tinnitus-specific CBT and personalized and frequency-matched sound therapy. Included patients presented to our tertiary clinic between 2017 and 2018, while those waitlisted were regarded as controls. Three surveys were administrated: Tinnitus Handicap Inventory (THI), Generalized Anxiety Disorder 7-item (GAD-7), and Perceived Stress Scale (PSS). A total of 30 patients enrolled in this study consisting of 20 treatment and 10 control patients and mean age was 55.4 ± 11.6 years. Treatment and control patients had similar age, sex, and pre-enrolment GAD and PSS (all P > .05). Baseline THI scores were also similar between treatment and control cohorts (50.1 ± 21.9 vs 62.0 ± 20.7; P = .15). After 8 weeks, though changes in GAD and PSS scores were similar (P > .05), the treatment group reported a significantly greater improvement in THI scores (17.7 ± 15.8 vs 5.3 ± 10.5, P = .04). This pilot study demonstrated potentially promising efficacy of a smartphone-based CBT and sound therapy platform for treating tinnitus and encourages future randomized controlled trials on this treatment modality.