Deep neural networks and image classification in biological vision.

In this paper we consider recent advances in the use of deep convolutional neural networks to understanding biological vision. We focus on claims about the plausibility of feedforward deep convolutional neural networks (fDCNNs) as models of image classification in the biological system. Despite the putative similarity of these networks to some properties of the biological vision system, and the remarkable levels of performance accuracy of some fDCNNs, we argue that their plausibility as a framework for understanding image classification remains unclear. We highlight two key issues that we suggest are relevant to the evaluation of any form of DNN used to examine biological vision: (1) Network transparency under analysis - that is, the challenge of understanding what networks do, and how they do it. (2) Identifying appropriate benchmarks for comparing network performance and the biological system using both quantitative and qualitative performance measures. We show that there are important divergences between fDCNNs and biological vision that reflect fundamental differences in computational architectures, and representational structures, supporting image classification in these networks and the biological system.

Study protocol for a randomised pilot study of a computer-based, non-pharmacological cognitive intervention for motor slowing and motor fatigue in Parkinson's disease.

BACKGROUND: Parkinson's disease (PD) is a chronic, neurodegenerative disorder affecting over 137,000 people in the UK and an estimated five million people worldwide. Treatment typically involves long-term dopaminergic therapy, which improves motor symptoms, but is associated with dose-limiting side effects. Developing effective complementary, non-pharmacological interventions is of considerable importance. This paper presents the protocol for a three-arm pilot study to test the implementation of computer-based cognitive training that aims to produce improvements or maintenance of motor slower and motor fatigue symptoms in people with PD. The primary objective is to assess recruitment success and usability of external data capture devices during the intervention. The secondary objectives are to obtain estimates of variance and effect size for changes in primary and secondary outcome measures to inform sample size calculations and study design for a larger scale trial. METHODS: The study aims to recruit between 40 and 60 adults with early- to middle-stage PD (Hoehn and Yahr 1-3) from National Health Service (NHS) outpatients' clinics and support groups across North Wales, UK. Participants will be randomised to receive training over five sessions in either a spatial grid navigation task, a sequential subtraction task or a spatial memory task. Patient-centred outcome measures will include motor examination scores from part 3 of the UPDRS-III and data from movement kinematic and finger tapping tasks. DISCUSSION: The results of this study will provide information regarding the feasibility of conducting a larger randomised control trial of non-pharmacological cognitive interventions of motor symptoms in PD. TRIAL REGISTRATION: ISRCTN, ISRCTN12565492. Registered 4 April 2018-retrospectively registered, in accordance with the WHO Trial Registration Data Set.

Eye movements during object recognition in visual agnosia.

This paper reports the first ever detailed study about eye movement patterns during single object recognition in visual agnosia. Eye movements were recorded in a patient with an integrative agnosic deficit during two recognition tasks: common object naming and novel object recognition memory. The patient showed normal directional biases in saccades and fixation dwell times in both tasks and was as likely as controls to fixate within object bounding contour regardless of recognition accuracy. In contrast, following initial saccades of similar amplitude to controls, the patient showed a bias for short saccades. In object naming, but not in recognition memory, the similarity of the spatial distributions of patient and control fixations was modulated by recognition accuracy. The study provides new evidence about how eye movements can be used to elucidate the functional impairments underlying object recognition deficits. We argue that the results reflect a breakdown in normal functional processes involved in the integration of shape information across object structure during the visual perception of shape.

Structure-based modulation of inhibition of return is triggered by object-internal but not occluding shape features.

When attention is oriented to an object it is inhibited from returning to the same object following a short delay. This inhibition-of-return (IOR) effect is modulated by an edge discontinuity presented between cue and target--an effect referred to as structure-based modulation of IOR. Here we examined two alternative accounts for the structure-based modulation effect. On one account the modulation is caused by the presence of any intervening feature between cue and target. On another account the modulation is caused by edge-bounded (i.e., closed) regions of space, on which space-based selection mechanisms operate. We presented cues and targets on unsegmented and internally or externally segmented rectangles to examine the two alternative accounts for the effect. Contrary to the predictions of the two alternative accounts, structure-based modulation of IOR was found with the internally but not with the externally segmented displays. This supports our hypothesis that object-based IOR arises from perceptually complete and internally structured object representations.