Functional Gradient Of The Fusiform Cortex For Chinese Character Recognition Part 3

RSA results

Semantic representations were only explored for RWs recognition, resulting in two clusters, the left middle and anterior FG, both of which were close to the lateral occipitotemporal sulcus (Fig. 3A).

The lateral occipitotemporal sulcus is an important part of the human brain. It is located on the side of the temporal lobe and has an inseparable relationship with human memory. The temporal lobe is one of the main areas in the human brain that processes language, cognition, hearing, memory, and other functions, and the lateral occipitotemporal sulcus is the most important part of this area.

Studies have found that the size of the lateral occipitotemporal sulcus is closely related to people's memory. The larger the occipital temporal sulcus, the better the person's memory. At the same time, the lateral occipitotemporal sulcus is also closely related to people's learning ability and information-processing abilities.

The size of the lateral occipitotemporal sulcus is largely genetically determined, but a person's lifestyle and environment can also affect the development of the lateral occipitotemporal sulcus. Maintaining good living habits and a healthy diet, maintaining a good mentality, and performing appropriate mental exercises are beneficial to the development of the occipital temporal sulcus.

In addition, human memory can be improved through continuous exercise and training. For example, we can improve our memory ability through continuous reading, memorizing, thinking, etc., thereby promoting the development of the lateral occipitotemporal sulcus.

In short, the lateral occipitotemporal sulcus is closely related to human memory. To maintain our memory, we need to pay attention to our living habits and environment and promote the development of the lateral occipitotemporal sulcus through continuous training to make our brains healthier, more flexible, and more flexible. wisdom. 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 10 ways to improve memory

Logo-grapheme representation is involved in the cognitive processing of orthography, radicals, and composite visual features, which were explored for FWs, PWs, and RWs recognition. 

FWs recognition did not include legal orthography and the logo-grapheme representations of FWs were found in the bilateral middle occipital gyrus (Fig. 3B). In addition to the bilateral middle and inferior occipital gyrus, the logo-grapheme representations of PWs were also found in the left middle FG (Fig. 3B), which may serve as abstract orthography because of the lack of semantics in PWs. 

The logo-grapheme representations of RWs were observed in the bilateral middle occipital gyrus, left inferior occipital gyrus, bilateral middle FG and left anterior FG (Fig. 3B). Particularly, the left middle and anterior FG were both involved in orthographic representations of RWs, but only the left middle FG participated in those of PWs, suggesting that the middle and anterior parts of the left FG have different functional roles. 

The left middle FG induced abstract orthography and the left anterior FG was related to lexical orthography. For more details, please see Table 3.

Notably, during RWs recognition, the logo-grapheme and semantic representations were observed in both the left middle and anterior FG and along with the lateral occipitotemporal sulcus. 

Clusters underlying logo-grapheme and semantic representations spatially neighbored each other in the left middle and anterior FG, respectively. To explore the relationships among the logo-grapheme and semantic representations between the left middle and anterior FG, Spearman's correlation analysis was conducted across subjects (Fig. 4B). 

A marginally significant correlation was found between the semantic representations in the left middle fusiform and left anterior fusiform regions (r = 0.26, p = 0.067). Logo-grapheme representations in the right middle FG were significantly correlated with the logo-grapheme representations in the left middle (r = 0.485, p, 0.001) and anterior FG (r = 0.325, p = 0.020). 

Logo-grapheme representations of the left anterior FG were significantly correlated with the semantic representations of the left middle FG (r = 0.284, p = 0.044). No significant correlation between the logo-grapheme representations in the left middle FG and left anterior FG was detected. 

Notably, as shown in Figure 4A, clusters underlying logo-grapheme and semantic representations in the left anterior FG and clusters in the left middle FG were neighbored or next to the anterior and posterior part of word form effect areas which were discovered during the activation analysis. However, for the lexical effect areas, no overlapping was found within the areas of logo-grapheme and semantic representations of RWs.

Validation results

As shown in Extended Data Figures 2-1, 3-1, 4-1 and Extended Data Tables 1-1, 2-1, 3-1, both the activation and RSA results based on data after excluding are consistent with the results based on all data, indicating that behavioral performance might have little effect on brain response of participants.

Discussion

In the current study, we aimed to investigate the functional gradient within the FG corresponding to different levels of orthographic structure in a visual lexical decision task to recognize four types of character-like stimuli. 

Different from the univariate analysis which identifies brain response to experimental stimulus through linearly fitting the behavior response with the hemodynamic activities of brain voxels, RSA characterizes the correspondence between brain activity patterns and theoretical/behavioral measurement (e.g., neural, and behavioral). 

Therefore, although both these two methods can characterize brain activities, RSA can detect more fine-grained pattern information than the univariate analysis. Our activation-based and RSA results revealed that there was a posterior-to-anterior gradient for the orthographic processing of character-like stimuli within the left FG. 

Besides, three functionally segregated regions within the left FG, a posterior, a middle, and an anterior region, were detected while no similar pattern was observed in the right FG. These findings revealed the neural basis for preprocessing of the hierarchical framework of Chinese orthography, i.e., general visual properties, radical orthography, orthography, and lexical orthography.

Functional gradients of character selectivity within the left FG

Although previous research revealed the involvement of the left FG in visual word recognition (Cohen et al., 2002; Bruno et al., 2008; Glezer et al., 2009, 2015; Baeck et al., 2015; Lochy et al., 2018), the levels of orthographic structure for the left FG involvement have not been elucidated (Kuo et al., 2004; Liu et al., 2008; Price and Devlin, 2011). 

Our results showed that the left occipitotemporal cortex preferentially responds to orthographically legal characters (i.e., RWs and PWs), which were consistent with previous findings (Price et al., 1996; Cohen et al., 2002; Ben-Shachar et al., 2007; Vinckier et al., 2007; Chan et al., 2009; Tian et al., 2020; Liu et al., 2021). Additionally, based on the minimum difference in orthographic legality between PWs and FWs, we found a wordform effect in the left middle FG, indicating the selectivity to orthographic legality, i.e., radical position for character identification (Wu et al., 2012). 

Furthermore, a lexical effect in the anterior part of the left FG was observed based on the minimum difference in lexical orthography between RWs and PWs, which indicated that the anterior part of the left FG may integrate phonological or semantic information from higher-level cortical areas such as the left angular gyrus, left supramarginal gyrus, and left inferior frontal gyrus, possibly through the arcuate fasciculus (Price et al., 2003; Liu et al., 2021).

Besides, we observed that PWs elicited more activations in the left middle FG, which were consistent with previous findings (Fiez et al., 1999; Xu et al., 2001). Meanwhile, FWs induced more activation in the posterior part of the left FG, while SCs elicited more activation in the left middle occipital gyrus. 

These findings support the prediction error hypothesis, which means that when a stimulus is recognized as potentially meaningful but is not predicted by its visual word form efficiently, it may elicit increased brain activity (Price and Devlin, 2011; J Zhao et al., 2019; Gagl et al., 2020). 

In line with previous findings in alphabetic languages, the varied activation patterns also revealed the corresponding relationship between functional gradient of the left FG and similarity to RWs, indicating the attuning to orthographic regularities of the reader's language in the course of learning to read (Vinckier et al., 2007).

Functional segregation of subregions in the left FG

To further examine the functional roles of the subregions of the left FG, we investigated the logo-grapheme representations of RWs, PWs, and FWs by using RSA methods. We observed that the logo-grapheme representations of RWs were detected in the middle and anterior parts of the left FG, whereas the logo-grapheme representations of PWs were only in the left middle FG, which might because of the difference between cognitive processing of RWs and PWs. These findings indicated that the left middle FG was processing word-form orthography, whereas the anterior part of the left FG was involved in lexical orthographic processing. 

Notably, in line with prior findings, we found that semantic representations in the left anterior FG and logo-grapheme representations in the left middle FG were well aligned with the anterior and posterior part of word form selective areas, respectively, indicating the functional subdivisions of left FG (Lerma-Usabiaga et al., 2018; White et al., 2019). 

Besides, the logo-grapheme representations of FWs were detected in the posterior region of the left FG. Therefore, despite highly discriminated linguistic features between Chinese and English (Mo et al., 2015), similar functional gradients of the left FG exist for both Chinese and alphabetic language processing, which indicates a radical-based stimulus scale in Chinese characters, like the letter-based stimulus scale in alphabetic languages (Vinckier et al., 2007; Lochy et al., 2018).

To identify the gradient of abstract orthography to lexical orthography from the middle part to the anterior part of the left FG, we also calculated the correlations between the brain representations of RWs. No significant correlation was found for logo-grapheme representations between the middle and anterior parts of the FG, which may imply that there are two different types of orthographic processing represented in the middle and anterior parts of the left FG. 

Meanwhile, a significant correlation between the logo-grapheme representations of the anterior part of the left FG and semantic representations of the middle part of the left FG was observed, which implied that the anterior region of the left FG might integrate semantic information from the left middle FG through top-down modulation to process orthography. 

Previous studies have revealed the existence of top-down modulation from high-level regions such as the left inferior frontal gyrus and left middle and superior temporal gyrus to the left middle FG (LB Zhao et al., 2017; Lerma-Usabiaga et al., 2018; Wang et al., 2018; Liu et al., 2021).

In general, both results of univariate activation analysis and RSA analysis confirmed functional gradients in the left FG but not the right FG during Chinese word recognition (Figs. 2, 3). Furthermore, RSA analysis provided more fine-grained results by voxel-wise decomposing cognitive components (logo-grapheme and semantics) of each task condition. 

Logo-grapheme representations and semantic representations of RWs in the left middle FG were included in the word-form effect area (Fig. 4A), which implied more than one cognitive process within a single functional gradient collectively supported its linguistic function. Potential associations between semantic representations in the left middle and anterior FG (Fig. 4B) showed possible interactions of cognitive components between different functional gradients. Future studies should focus on how functional gradients in the left FG are organized by investigating complex interactions of cognitive components within and between gradients.

Functional organization of character selectivity in the right FG

Because of the square shape of Chinese characters, substantial evidence has shown that the right FG is specifically involved in Chinese character recognition to process spatial information such as the locations of different strokes and radicals composing the character (Tan et al., 2000, 2001, 2005; Bolger et al., 2005; Guo and Burgund, 2010). 

We also found that not only real characters but pseudo-characters and false characters all elicited activation of the right FG. However, we did not find a hierarchical functional organization of Chinese orthography in the right FG, which was in line with previous findings (Vinckier et al., 2007; Chan et al., 2009; Kronschnabel et al., 2013; Olulade et al., 2013; LB Zhao et al., 2017; Tian et al., 2020). 

Given that the right FG was proposed to process radical configuration or visual-spatial information (Peyrin et al., 2006; Deng et al., 2011; Woodhead et al., 2011), character-like stimuli comprising strokes or radicals packed into a square shape may elicit similar activation patterns in the right FG. Additionally, it was indicated that the left FG stores information in terms of parts and their relationships to visual objects, whereas the right FG stores holistic information about visual objects (Dien, 2009).

Furthermore, neither a functional gradient of the logograph eme representations for character-like stimuli nor semantic representation were found for the right FG, which may indicate that the right FG was only involved in visual-spatial processing rather than lexical processing during Chinese character recognition. Notably, we found significant correlations between the logo-grapheme representation of RWs in the right middle FG and that of RWs in the left middle and anterior FG. 

Several lesion studies have proposed that the splenium of the corpus callosum links the left FG to its right homolog, thereby integrating visual information projected to bilateral visual areas (Binder and Mohr, 1992; Molko et al., 2002; Shan et al., 2010). Our results suggested that the orthographic representations of Chinese characters may integrate visual-spatial information from the right middle FG and orthographic information from the left FG.

Two limitations of this study should be addressed. First, although we speculated that the anterior region of left FG may receive top-down modulation from higher-level brain regions such as the left inferior frontal gyrus and left superior and middle temporal gyrus, the present study could not provide direct evidence for this implication because of the limitations of the temporal resolution of fMRI. 

Future studies employing other imaging methods should be conducted to test this assumption. In addition, a recent intracranial recordings study has suggested that functional gradient within the left FG may represent varying degrees of top-down influence from the left middle FG to the primary visual cortex (Woolnough et al., 2021), which further emphasizes the importance of multiple modality studies in the future. 

Second, our data cannot determine whether subregions within the fusiform cortex are involved in bottom-up only or interactive bottom-up and top-down processes, as stated by two of the main theoretical proposals regarding the functional role of this region. Future studies exploring the interactions among orthography and higher level linguistic processes would be helpful for this question (i.e., phonology and semantics).

In conclusion, we observed a posterior-to-anterior functional gradient of character-like stimuli with increasing sensitivity from SCs to real characters within the left fusiform cortex but not in its right homolog. Based on RSA results, we identified that the left middle FG was involved in word form orthographic processing, while the anterior part of the left FG was involved in lexical orthographic processing. 

These findings indicated that the left fusiform cortex presents a posterior-to-anterior gradient corresponding to the lower-to-higher likeness of character type during Chinese character recognition.

References

Allen M, Poggiali D, Whitaker K, Marshall TR, Kievit RA (2019) Raincloud plots: a multi-platform tool for robust data visualization. Wellcome Open Res 4:63.

Baeck A, Kravitz D, Baker C, Op de Beeck HP (2015) Influence of lexical status and orthographic similarity on the multi-voxel response of the visual word form area. Neuroimage 111:321–328.

Baker CI, Hutchison TL, Kanwisher N (2007) Does the fusiform face area contain subregions highly selective for nonfaces? Nat Neurosci 10:3–4.

Ben-Shachar M, Dougherty RF, Deutsch GK, Wandell BA (2007) Differential sensitivity to words and shapes in ventral occipitotemporal cortex. Cereb Cortex 17:1604–1611.

Binder JR, Mohr JP (1992) The topography of callosal reading pathways - a case-control analysis. Brain 115:1807–1826.

Binder JR, Medler DA, Westbury CF, Liebenthal E, Buchanan L (2006) Tuning of the human left fusiform gyrus to the sublexical orthographic structure. Neuroimage 33:739–748.

Bolger DJ, Perfetti CA, Schneider W (2005) Cross-cultural effect on the brain revisited: universal structures plus writing system variation. Hum Brain Mapp 25:92–104.

Bruno JL, Zumberge A, Manis FR, Lu ZL, Goldman JG (2008) Sensitivity to orthographic familiarity in the occipitotemporal region. Neuroimage 39:1988–2001.

Centanni TM, King LW, Eddy MD, Whitfield-Gabrieli S, Gabrieli JDE (2017) Development of sensitivity versus specificity for print in the visual word form area. Brain Lang 170:62–70.

Chan ST, Tang SW, Tang KW, Lee WK, Lo SS, Kwong KK (2009) Hierarchical coding of characters in the ventral and dorsal visual streams of Chinese language processing. Neuroimage 48:423– 435.

Cohen L, Lehéricy S, Chochon F, Lemer C, Rivaud S, Dehaene S (2002) Language-specific tuning of visual cortex functional properties of the visual word form area. Brain 125:1054–1069.

Coltheart M, Rastle K, Perry C, Langdon R, Ziegler J (2001) DRC: a dual route cascaded model of visual word recognition and reading aloud. Psychol Rev 108:204–256.

For more information:1950477648nn@gamil.com