Long flankers alleviate visual crowding by enhancing population target coding in macaque V1

Visual crowding refers to impaired recognition of stimuli in cluttered scenes and is often regarded as a fundamental limitation of object recognition in peripheral vision.^1,2 Recent evidence suggests that crowding can be weakened or even abolished when the flanker configuration is changed.^3,4,5,6,7,8 For example, crowding is substantially reduced when a target line is flanked by long lines rather than by equal-length lines, producing an "uncrowding" effect.^4,5,7 However, the neural mechanisms underlying this uncrowding effect remain unclear. Here, we used two-photon calcium imaging to test the hypothesis that long flankers reduce crowding by alleviating surround suppression in V1 and/or V4. We simultaneously recorded responses of large populations of V1 or V4 neurons to a line target flanked by either equal-length or long lines in awake, fixating macaques. Compared with equal-length flankers, long flankers alleviated surround suppression in V1 by reducing average response suppression and improving orientation representational geometry and decoding accuracy to levels comparable to the baseline target-alone condition. In contrast, long flankers in V4 produced stronger suppression and further degraded representational geometry and decoding accuracy. These results indicate that long flankers reduce surround suppression and crowding in V1, with their signals potentially bypassing V4 during readout.