关键词: 视觉掩蔽, 视觉搜索, 近红外脑功能成像, 顶枕叶

Abstract:

When our visual system processes target signals, it usually receives large amounts of irrelevant information from the target, leading to a reduction in the visibility of the target. A wealth of research has shown that visual search for target letters against a masking background is largely determined by the masker type. Informational maskers, such as either randomly positioned and oriented letters or randomly distributed letter fragments, induce stronger masking effects on recognition of target letters than the energetic maskers do, such as the random-phase masker (same spectral amplitude composition as the letter masker but with the phase spectrum randomized) or the random-pixel masker (the locations of the letter maskers’ pixel amplitudes being randomized). However, the mechanisms under informational masking and those under energetic masking are still unknown.

The current study examined both cortical activities and behavioral performances in the visual search task, which is determined by whether one of four letters presented at four symmetrically-located positions differs from the others under three masking conditions (random pixels, letter fragments, and random letters). Both the oxygenated hemoglobin concentration (HbO) responses in the primary visual cortex (V1) and secondary visual cortex (V2) with a functional near infrared spectroscopy (fNIRS) were recorded. Twenty (4 males, 16 females) healthy adults (mean age: 22.5 ± 1.67 years) participated in the experiment. Each masking condition contained 5 blocks, and each block contained 8 trails. There was a resting phase of 20 seconds between the two blocks. Spatial registration methods were applied to localize the cortical regions underneath each channel and to define two regions of interest (ROIs), which are the primary visual cortex (V1) and secondary visual cortex (V2).

The behavioral results showed that the performance of recognizing target letters improved when the masker type shifted from random letters to letter fragments and to random pixels, suggesting that the letter masker interfered the most with performance than the letter fragment and random-pixel maskers. The random-pixel masker caused the least masking effect. The fNIRS results showed that both letter masker and letter-fragment masker produced an increase in cortical oxygen level. Many regions of interest (ROIs), particularly the visual cortex (including V1 and V2), were more activated under the letter or the letter-fragment masking condition compared to the random-pixel masking condition. Moreover, the differences in cortical activation between the masking conditions further suggested that the V1 and V2 are the critical brain regions involved in visual letter search and informational masking of letter recognition.

To summarize, this study used fNIRS to explore the cortex activation patterns of different types of masking on target recognition. The results showed that information masking had much more interference on visual search and caused greater processing loads in primary and secondary visual cortex, compared with energy masking under the same conditions. Furthermore, the differences between letter fragments masking and letters masking are reflected in the activation mode of V1 and V2 regions.

Key words: visual masking, visual research, functional near infrared spectroscopy (fNIRS), parietal-occipital cortices