Functional Gradient Of The Fusiform Cortex For Chinese Character Recognition Part 1

Abstract

Visual word recognition has been proposed to have a functional and spatial organization corresponding to hierarchical language-like word forms in the left fusiform gyrus (FG) during visual word recognition in alphabetic languages. 

Visual word recognition and memory are closely related. Research shows that training visual word recognition skills can improve the brain's memory ability and promote the overall development of brain function.

First, visual word recognition improves our working memory. Working memory is a form of short-term memory that helps us process and store information while performing complex cognitive tasks. For example, if we want to read a book, we need to remember multiple information such as the content of the book, page number, and chapter at the same time, which requires us to have a strong working memory ability. Visual word recognition training can enhance our ability to process and remember text information, and effectively improve our working memory level.

Secondly, visual word recognition ability can also improve our long-term memory ability. Long-term memory refers to information that is stored in the brain and has strong persistence, including language, historical events, cultural knowledge, etc. Visual word recognition training can help us deepen our understanding and memory of text information, thereby promoting the formation and consolidation of long-term memory.

In short, visual word recognition ability plays an important role in promoting our memory. By training visual word recognition skills, we can improve our working memory and long-term memory capabilities, laying a solid foundation for future study and work. Let us be positive, continue to explore the mysteries of visual word recognition training, and improve our cognitive level. It can be seen that we need to improve memory, and Cistanche deserticola can significantly improve memory because Cistanche deserticola is a traditional Chinese medicinal material that has many unique effects, one of which is to improve memory. The efficacy of minced meat comes from the various active ingredients it contains, including acid, polysaccharides, flavonoids, etc. These ingredients can promote brain health in various ways.

Click Know to improve short-term memory

However, it is still unclear whether similar functional gradients of word-like representation exist during Chinese character recognition. 

In this study, we adopted univariate activation analysis and representational similarity analysis (RSA) methods to investigate the functional organization in the FG for Chinese character recognition using task fMRI data. Native Chinese readers were visually presented with four types of character-like stimuli (i.e., real characters, pseudo-characters, false characters, and stroke combinations). 

After analysis, we observed a posterior-to-anterior functional gradient in the left FG corresponding to the degree of likeness of stimuli to character. Additionally, distinct subregions of the left FG harbor different orthographic codes. The middle part of the left FG was involved in abstract orthographic processing, while the anterior part of the left FG was involved in lexical orthographic processing (i.e., mapping orthography onto phonology or semantics). 

Notably, for the right FG, we did not find a similar coding pattern for selectivity to character likeness, indicating the asymmetry of the functional hierarchical organization in favor of the left hemisphere. 

In conclusion, our findings revealed that the left FG presents a posterior-to-anterior gradient functional processing for Chinese character recognition, which expands our understanding of the psychological, neural, and computational theories of word reading.

Keywords: 

Chinese character recognition; functional gradient; fusiform cortex; representational similarity analysis; univariate activation analysis.

Significance Statement

The left fusiform gyrus (FG) is essential to reading, yet its functional organization during Chinese reading remains unclear. 

Here, we revealed a posterior-to-anterior functional gradient corresponding to the lower to higher character-like stimuli within the left FG during Chinese character recognition but not in its right homolog. 

Employing representational similarity analysis (RSA), we identified two functionally segregated subregions of the left FG: the middle part for word-form orthographic processing and the anterior part for lexical orthographic processing. 

In conclusion, we found the posterior, middle, and anterior regions of the left FG are responsive to distinct orthographic hierarchies thereby performing different but complementary computations. Based on this gradient pattern, the left FG interacts with other regions of the language network to achieve Chinese reading.

Introduction

Efficient visual word recognition requires a fast conversion of word form and orthography to word pronunciation and meaning (Liu, 1999; Coltheart et al., 2001; Price and Devlin, 2011). 

Neuroimaging and lesion studies have revealed that the left fusiform gyrus (FG) is critical for such conversion during word reading (Kuo et al., 2001; Cohen et al., 2002; Baker et al., 2007; Dehaene et al., 2010; Centanni et al., 2017). 

Additionally, the lateral middle region of the left FG called the visual word form area, is thought to be spatially reproducible across different writing systems that vary greatly in the type of scripts, such as alphabetic languages (e.g., English) and logographic languages (e.g., Chinese characters; Bolger et al., 2005; Liu et al., 2008; Dehaene and Cohen, 2011). 

A functional hierarchical organization of word-like stimuli within the left FG during English word reading has been observed (Vinckier et al., 2007). However, whether a similar internal organization of the left FG exists in Chinese word reading is still unclear.

Recently, some studies have examined the functional organization of word-like stimuli in the ventral occipitotemporal cortex (vOT). For alphabetic languages, lines of evidence based on activation results have observed a functional and spatial hierarchical organization in the left FG during visual word recognition (Binder et al., 2006; Vinckier et al., 2007; Van der Mark et al., 2009; Kronschnabel et al., 2013; Olulade et al., 2013, 2015; Lerma-Usabiaga et al., 2018). 

Vinckier and colleagues found that different levels of orthographic stimuli induced equal activation in the posterior part of the left FG, whereas more word-like stimuli induced higher activation along the middle to the anterior axis (Vinckier et al., 2007). 

Consistently, an intracranial recording study on English word recognition confirmed that the posterior part of the left FG was uniquely involved in letter selectivity, but emphasized the spatially intermingled but not strict hierarchical organization underlying prelexical and lexical responses in the middle and anterior regions of the left FG (Lochy et al., 2018). 

Those authors consistently identified that for the left FG, the posterior part was involved in letter processing and emphasized the functional gradient from the middle to the anterior part.

Given the sharp difference between written English and Chinese in orthographic structure, two recent studies have investigated whether a similar functional gradient of brain activity for character-like stimuli exists in Chinese (Chan et al., 2009; Tian et al., 2020). 

Chan and colleagues found that the anterior region of the left FG was more selective for Chinese character-like stimuli with orthographic legality, whereas the posterior part was more selective for Korean characters (Chan et al., 2009). Tian and colleagues suggested that the anterior and middle regions of the left FG were more selective for radical-based stimuli, whereas the posterior region was not (Tian et al., 2020). 

However, the corresponding relationship between different levels of Chinese orthographic structure to subregions of the left FG has still not been revealed. In addition, the right FG was also significantly activated, which was interpreted as spatial information processing during Chinese word recognition (Tan et al., 2000, 2001). 

However, which levels of orthography were processed and whether divergent hierarchical coding patterns existed in the right FG during Chinese word reading also remained largely unknown.

The current study examined the functional organization in the FG during Chinese character recognition by using univariate activation analysis and RSA methods. 

Here, we recruited a group of adults, native Chinese speakers who performed a lexical decision task for real words (RWs), pseudowords (PWs), false words (FWs), and stroke combinations (SCs) during fMRI scanning. 

Given that Chinese orthographic processing entails four main components: visual properties, radical orthography, word-form orthography, and lexical orthography, we hypothesize that distinct components take place in distinct subregions of the left FG, which resulting in a posterior-to-anterior gradient of Chinese orthographic processing.

Materials and Methods

Participants

Fifty-one college students (mean age = 23.4 years, 19–28 years old, 25 males/26 females) were recruited in the current study by online advertising. All were native Chinese speakers with normal or corrected-to-normal vision over 4.8 (Logarithmic Vision Chart Values). 

Forty-one were identified as right-handed, and the rest had balanced handedness according to the Edinburgh Handedness Inventory (Oldfield, 1971). None had any history of neurological disease or psychiatric disorders. 

Informed written consent was provided to each subject before the experiment. The current study was approved by the Ethics Committee of the School of Life Sciences, Fudan University.

Stimuli and task fMRI procedures

The stimuli set consisted of four conditions: RWs, PWs, FWs, and SCs, with 40 trials in each condition (Fig. 1A). Chinese orthographic processing entails processing four putative components, that is, visual properties, radical orthography, word-form orthography, and lexical orthography, which construct a hierarchical framework of cognitive processes (Fig. 1B). 

RWs are high-frequency single-character words consisting of two radicals. PWs are formed by two radicals that are presented at their legal positions but cannot be found in the existing Chinese dictionary. Notably, in contrast with PWs in alphabetic language, PWs in Chinese are both unpronounceable and meaningless, even without phonological and semantic cues. 

FWs are formed by two radicals presented in illegal positions. SCs are comprised of randomly arranged strokes that appear in real characters and maintain the same envelope as real characters. 

The horizontal visual angle of all stimuli, which were white and presented on a black screen, was 4.37°. The percentage of pixels, picture size, and number of strokes were matched across conditions. The word frequency of RWs and single-character words used to build PWs and FWs were also matched.

In the current study, an event-related design and lexical decision task were adopted. Each stimulus was presented for 600 ms in randomized order, with a randomized interstimulus interval (ISI) ranging from 4000 to 6000 ms. 

A fixation cross was presented in the center of the screen during ISI to obtain baseline brain activity (Fig. 1A). The lexical decision task required participants to judge whether the stimulus was a real character by pressing buttons with their right index fingers. Notably, the criterion for identifying a real character was whether it has meaning or not. 

A practice section consisting of 16 trials (an additional four stimuli in each condition) was conducted out of the scanner before the normal experiment to ensure a full understanding of task demands.

fMRI acquisition and data preprocessing

Functional and structural magnetic resonance imaging data were collected by a 3.0-T Siemens Prisma scanner with a 32-channel head coil (Siemens Healthcare) at Zhangjiang International Brain Imaging Center (ZIC) of Fudan University, Shanghai, China. 

An echo planar imaging (EPI) sequence was used for functional imaging acquisition [TR = 720 ms, TE = 33 ms, flip angle = 52°, matrix size = 110
96, field of view (FOV) = 220
196 mm, slice thickness = 2 mm, number of slices = 72]. 

Anatomical, high-resolution, T1-weighted images were collected before tasks (TR =3000ms, TE = 2.56ms, flip angle = 8°, matrix size = 320
320, FOV = 256
256 mm, slice thickness = 0.8 mm, number of slices = 208).

Image preprocessing was conducted by Statistical Parametric Mapping-12 (SPM12, Wellcome Trust Centre for Neuroimaging, London, United Kingdom; http://www. fil.ion.ucl.ac.uk/spam). First, several volumes were not recorded before the trigger launch to ensure T1 equilibrium. 

Volumes were temporally realigned to middle EPI volume and spatially realigned to correct head movement. The structural image of each subject was registered to the mean EPI image, segmented, and normalized to Montreal Neurologic Institute (MNI) space. 

The realigned EPI volumes were normalized to MNI space by deformation field parameters from structural image normalization. The normalized EPI volumes were smoothed with a 6 mm Gaussian kernel and high-pass filter.

For more information:1950477648nn@gmail.com