BackgroundHomonymous visual field deficit is a common clinical problem, most frequently caused by stroke or traumatic brain injury. After 3 months, the visual field loss is generally considered to be permanent. In certain cases, however, information within the blind hemifield can influence behaviour despite the individual often having no conscious awareness of the stimulus. The mechanism and extent of this function remains poorly understood, especially in cases of adult-onset injury. We aimed to investigate the ability of patients with primary visual cortex damage sustained in adulthood to detect and discriminate direction of low-luminance moving targets within the blind visual field. Our hypothesis was that individuals will retain some ability beyond chance, optimally for speeds between 8°/s and 30°/s. MethodsParticipants were recruited from ophthalmological or stroke services in three UK centres. Testing was done at the John Radcliffe Hospital, Oxford. Of 14 participants recruited, two were excluded after structural imaging showed almost complete damage to the occipital lobe, extending to parietal and temporal lobes including subcortical nuclei. Cases with additional non-correctable visual impairment, or previous neurological disease, were also excluded. All patients had sustained unilateral damage to the primary visual cortex (V1), causing homonymous visual field loss recorded by Humphrey perimetry. Pathological changes had been caused by posterior circulation stroke in 11 patients, or by benign tumour resection in three patients, at least 6 months previously. Psychophysical testing was conducted with a 60Hz monitor at a distance of 68 cm. Visual stimuli consisted of moving black dots (8 dots per deg2) in an aperture of 5° or 8° diameter. Stimulus location was restricted to the scotoma, on a grey background. Each stimulus appeared for 500 ms with jittered onset while the participant fixated on a central black cross. In experiment 1, participants indicated whether a stimulus appeared in the first or second time-interval. If they saw nothing, they were instructed to guess. In experiment 2, participants reported horizontal or vertical direction of motion. In both experiments, stimulus speed was altered parametrically: 4°/s, 8°/s, 20°/s, 32°/s, with 20 trials per condition. Fixation was recorded with an eye tracker, and any trials containing eye movements were removed from analysis. All patients underwent high-resolution structural MRI on the day of testing. Primary outcome was whether patients could demonstrate significant detection or discrimination of visual stimuli above chance. Secondary analysis considered how accuracy altered as a function of stimulus speed. Statistical testing was modelled with a binomial distribution. Written and verbal consent was given by all participants. Ethics approval was provided by the Oxford Research Ethics Committee (ref B 08/H0605/156). FindingsEight patients showed significant detection above chance (patient [P] 1 94%, p<0·01; P2 55%, p=0·04; P3 57%, p=0·03; P4 58·8%, p=0·02; P5 55·8%, p=0·04; P6 67·3%, p=0·01; P7 69·9%, p=0·001; P8 100%, p<0·001), and three patients correctly discriminated direction of motion (P4 75%, p=0001; P6 60%, p=0·018; P7 58·7%, p=0·026). Group analysis revealed an inverted-U pattern of accuracy as a function of speed across both experiments. Performance was best at 20°/s (mean detection 60·3% [SE 3·1], p<0·001, with patients 1 and 8 excluded from analysis because of performance at ceiling; mean discrimination 56·0% [SE 2·6], p=0·01), and also significantly above chance at 4°/s and 8°/s, but not 32°/s. InterpretationThis study is one of the few to assess visual performance in patients with homonymous visual field deficit due to primary visual cortex damage in adulthood. We have shown that even low-contrast complex motion can elicit blindsight in most patients, with optimum detection and discrimination at speeds of 20°/s. In some cases, particularly for direction discrimination, differences in performance can be subtle. Such trends become more apparent at a group level, or perhaps if patients were to undertake a greater number of trials. Performance as a function of speed echoes research into motion processing in area MT+/V5, where an optimum response is seen for intermediate speeds. It provides further evidence for a direct, perhaps subcortical route to this region when V1 is damaged, which may be present in most patients with these areas intact. That slow motion (postulated to depend upon V1 processing) was discriminated above chance, though small, warrants further investigation. FundingRoyal Society, Wellcome Trust, NIHR Oxford Biomedical Research Centre.