Assessing Vision Quality in Retinal Prosthesis Implantees through Deep Learning: Current Progress and Improvements by Optimizing Hardware Design Parameters and Rehabilitation.

Retinal prosthesis (RP) is used to partially restore vision in patients with degenerative retinal diseases. Assessing the quality of RP-acquired (i.e., prosthetic) vision is needed to evaluate RP impact and prospects. Spatial distortions caused by electrical stimulation of the retina in RP, and the low number of electrodes, have limited the prosthetic vision: patients mostly localize shapes and shadows rather than recognizing objects. We simulate prosthetic vision and evaluate vision on image classification tasks, varying critical hardware parameters: total number and size of electrodes. We also simulate rehabilitation by re-training our models on prosthetic vision images. We find that electrode size has little impact on vision while at least 400 electrodes are needed to sufficiently restore vision (more than 65% classification accuracy on a complex visual task after rehabilitation). Argus II, a currently available implant, produces a low-resolution vision leading to low accuracy (21.3% score after rehabilitation) in complex vision tasks. Rehabilitation produces significant improvements (accuracy improvement of up to 30% on complex tasks, depending on the number of electrodes) in the attained vision, boosting our expectations for RP interventions and motivating the establishment of rehabilitation procedures for RP implantees.

Using Biologically-inspired Image Features to Model Retinal Response: Evidence from Biological Datasets.

Retinal models are needed to simulate the translation of visual percepts to Retinal Ganglion Cells (RGCs) neural spike trains, through which visual information is transmitted to the brain. Restoring vision through neural prostheses motivates the development of accurate retinal models. We integrate biologically-inspired image features to RGC models. We trained Linear-Nonlinear models using response data from biological retinae. We show that augmenting raw image input with retina-inspired image features leads to performance improvements: in a smaller (30sec. of retina recordings) set integration of features leads to improved models in approximately $\frac{2}{3}$ of the modeled RGCS; in a larger (4min. recording) we show that utilizing Spike Triggered Average analysis to localize RGCs in input images and extract features in a cell-based manner leads to improved models in all (except two) of the modeled RGCs.

Stratification of carotid atheromatous plaque using interpretable deep learning methods on B-mode ultrasound images.

Carotid atherosclerosis is the major cause of ischemic stroke resulting in significant rates of mortality and disability annually. Early diagnosis of such cases is of great importance, since it enables clinicians to apply a more effective treatment strategy. This paper introduces an interpretable classification approach of carotid ultrasound images for the risk assessment and stratification of patients with carotid atheromatous plaque. To address the highly imbalanced distribution of patients between the symptomatic and asymptomatic classes (16 vs 58, respectively), an ensemble learning scheme based on a sub-sampling approach was applied along with a two-phase, cost-sensitive strategy of learning, that uses the original and a resampled data set. Convolutional Neural Networks (CNNs) were utilized for building the primary models of the ensemble. A six-layer deep CNN was used to automatically extract features from the images, followed by a classification stage of two fully connected layers. The obtained results (Area Under the ROC Curve (AUC): 73%, sensitivity: 75%, specificity: 70%) indicate that the proposed approach achieved acceptable discrimination performance. Finally, interpretability methods were applied on the model's predictions in order to reveal insights on the model's decision process as well as to enable the identification of novel image biomarkers for the stratification of patients with carotid atheromatous plaque.Clinical Relevance-The integration of interpretability methods with deep learning strategies can facilitate the identification of novel ultrasound image biomarkers for the stratification of patients with carotid atheromatous plaque.

Machine Learning Method for Functional Assessment of Retinal Models.

Challenges in the field of retinal prostheses motivate the development of retinal models to accurately simulate Retinal Ganglion Cells (RGCs) responses. The goal of retinal prostheses is to enable blind individuals to solve complex, reallife visual tasks. In this paper, we introduce the functional assessment (FA) of retinal models, which describes the concept of evaluating the performance of retinal models on visual understanding tasks. We present a machine learning method for FA: we feed traditional machine learning classifiers with RGC responses generated by retinal models, to solve object and digit recognition tasks (CIFAR-10, MNIST, Fashion MNIST, Imagenette). We examined critical FA aspects, including how the performance of FA depends on the task, how to optimally feed RGC responses to the classifiers and how the number of output neurons correlates with the model's accuracy. To increase the number of output neurons, we manipulated input images - by splitting and then feeding them to the retinal model and we found that image splitting does not significantly improve the model's accuracy. We also show that differences in the structure of datasets result in largely divergent performance of the retinal model (MNIST and Fashion MNIST exceeded 80% accuracy, while CIFAR-10 and Imagenette achieved ∼40%). Furthermore, retinal models which perform better in standard evaluation, i.e. more accurately predict RGC response, perform better in FA as well. However, unlike standard evaluation, FA results can be straightforwardly interpreted in the context of comparing the quality of visual perception.

New-Onset Psychosis Associated With a Lesion Localized in the Rostral Tectum: Insights Into Pathway-Specific Connectivity Disrupted in Psychosis.

OBJECTIVE: To investigate pathway-specific connectivity disrupted in psychosis. METHODS: We carried out a case study of a middle-aged patient who presented with new-onset psychosis associated with a space-occupying lesion localized in the right superior colliculus/periaqueductal gray. The study sought to investigate potential connectivity deficits related to the lesion by the use of diffusion tensor imaging and resting-state functional magnetic resonance imaging. To this aim, we generated a functional connectivity map of the patient's brain, centered on the lesion area, and compared this map with the corresponding map of 10 sex- and age-matched control individuals identified from the Max Planck Institute-Leipzig Mind-Brain-Body database. RESULTS: Our analysis revealed a discrete area in the right rostral tectum, in the immediate vicinity of the lesion, whose activity is inversely correlated with the activity of left amygdala, whereas left amygdala is functionally associated with select areas of the temporal, parietal, and occipital lobes. Based on a comparative analysis of the patient with 10 control individuals, the lesion has impacted on the connectivity of rostral tectum (superior colliculus/periaqueductal gray) with left amygdala as well as on the connectivity of left amygdala with subcortical and cortical areas. CONCLUSIONS: The superior colliculus/periaqueductal gray might play important roles in the initiation and perpetuation of psychosis, at least partially through dysregulation of left amygdala activity.

Biologically-inspired image processing in computational retina models.

Retinal Prosthesis (RP) is an approach to restore vision, using an implanted device to electrically stimulate the retina. A fundamental problem in RP is to translate the visual scene to retina neural spike patterns, mimicking the computations normally done by retina neural circuits. Towards the perspective of improved RP interventions, we propose a Computer Vision (CV) image preprocessing method based on Retinal Ganglion Cells functions and then use the method to reproduce retina output with a standard Generalized Integrate & Fire (GIF) neuron model. "Virtual Retina" simulation software is used to provide the stimulus-retina response data to train and test our model. We use a sequence of natural images as model input and show that models using the proposed CV image preprocessing outperform models using raw image intensity (interspike-interval distance 0.17 vs 0.27). This result is aligned with our hypothesis that raw image intensity is an improper image representation for Retinal Ganglion Cells response prediction.