98%
921
2 minutes
20
The ganglion cells of primate retina have center-surround receptive fields. A strong candidate for mediating linear surround circuitry is negative feedback from the H1 horizontal cell onto the cone pedicle. We measured the spatial properties of H1 cell receptive fields in the in vitro macaque monkey retina using sinusoidal gratings, spots, and annuli. Spatial tuning curves ranged in shape from smoothly low pass to prominently notched. The tuning curves of approximately 80% of cells could be well described by a sum of two exponentials, giving a prominent central peak superimposed on a broad shallow skirt. The mean diameter of the combined receptive field decreased with eccentricity from 309 micro m at 11 mm to 122 micro m at 4 mm. We propose that the strong narrow field reflects direct synaptic input from the cones overlying the dendritic tree whereas the weak wide field reflects coupled inputs from neighboring H1 cells. Those cells not well fit by a sum of exponentials had tuning curves with additional peaks at higher spatial frequencies that were likely due to undersampling in the cone-H1 network. Unlike other vertebrates, the macaque H1 network is less strongly coupled, has smaller receptive fields, and shows no functional plasticity. Macaque H1 receptive fields are surprisingly small, suggesting a great reduction in electrical coupling. Because the center of the H1 receptive field gets only a small percentage of its total response from the coupled field, the smallest receptive fields are similar in diameter to the dendritic trees. They are probably small enough to form the surrounds of foveal midget cells. The H1 network is compatible with a mixed-surround model of spectral opponency.
Download full-text PDF |
Source |
---|---|
http://dx.doi.org/10.1167/2.4.1 | DOI Listing |
Med Phys
September 2025
School of Computer, Electronics and Information, Guangxi University, Nanning, China.
Background: Deformable medical image registration is a critical task in medical imaging-assisted diagnosis and treatment. In recent years, medical image registration methods based on deep learning have made significant success by leveraging prior knowledge, and the registration accuracy and computational efficiency have been greatly improved. Models based on Transformers have achieved better performance than convolutional neural network methods (ConvNet) in image registration.
View Article and Find Full Text PDFCommun Biol
September 2025
Department of Physiology Anatomy and Genetics, University of Oxford, Oxford, UK.
Primate lateral intraparietal area (LIP) has been directly linked to perceptual categorization and decision-making. However, the intrinsic LIP circuitry that gives rise to the flexible generation of motor responses to sensory instruction remains unclear. Using retrograde tracers, we delineate two distinct operational compartments based on different intrinsic connectivity patterns of dorsal and ventral LIP.
View Article and Find Full Text PDFIEEE Trans Neural Netw Learn Syst
September 2025
In industrial scenarios, semantic segmentation of surface defects is vital for identifying, localizing, and delineating defects. However, new defect types constantly emerge with product iterations or process updates. Existing defect segmentation models lack incremental learning capabilities, and direct fine-tuning (FT) often leads to catastrophic forgetting.
View Article and Find Full Text PDFCereb Cortex
August 2025
Nencki Institute of Experimental Biology, PAS, 3 Pasteur Street, 02-093 Warsaw, Poland.
In the visual cortices, receptive fields (RFs) are arranged in a gradient from small sizes in the center of the visual field to the largest sizes at the periphery. Using functional magnetic resonance imaging (fMRI) mapping of population RFs, we investigated RF adaptation in V1, V2, and V3 in patients after long-term photoreceptor degeneration affecting the central (Stargardt disease [STGD]) and peripheral (Retinitis Pigmentosa [RP]) regions of the retina. In controls, we temporarily limited the visual field to the central 10° to model peripheral loss.
View Article and Find Full Text PDFPLoS One
September 2025
School of Medical Engineering, Xinxiang Medical University, Xinxiang, China.
Computer-aided diagnostic (CAD) systems for color fundus images play a critical role in the early detection of fundus diseases, including diabetes, hypertension, and cerebrovascular disorders. Although deep learning has substantially advanced automatic segmentation techniques in this field, several challenges persist, such as limited labeled datasets, significant structural variations in blood vessels, and persistent dataset discrepancies, which continue to hinder progress. These challenges lead to inconsistent segmentation performance, particularly for small vessels and branch regions.
View Article and Find Full Text PDF