Skip to main content
NCBI home page
Springer logoLink to Springer
. 2026 Mar 28;105(4):210. doi: 10.1007/s00277-026-06893-6

Aberrant functional connectivity and Granger causality analysis in patients with β-thalassemia major

Haifeng Zheng 1,#, Hang Yu 2,#, Xingye Yang 1,#, Xi Deng 1, Meiru Bu 1,3, Yuhong Qin 1, Bihong T Chen 4, Yifeng Wang 2,✉, Muliang Jiang 1,5,✉
PMCID: PMC13033005  PMID: 41902954

Abstract

This study aimed to assess alterations in functional connectivity (FC) within brain networks in children and adolescents with β-TM major and to explore the intrinsic relationship between network changes and cognitive impairment. This prospective study recruited 70 patients with β-TM and 64 healthy controls. Cognitive function assessments using Montreal Cognitive Assessment (MoCA) and Modified Mini-Mental State Examination (MMMSE), and hematological parameters were collected. Region of interest (ROI)-to-ROI connectivity analysis was conducted to investigate the whole brain FC within and between resting-state networks. Granger causality analysis was utilized to evaluate the effective interactions among them. Patients exhibited significant cognitive impairment compared to controls. Key hematological indicators, such as serum ferritin, were not found to be correlated with cognitive function. Rs-fMRI revealed extensive reductions in functional connectivity, accompanied by several enhancements. These FCs’ alterations significantly correlated with cognitive deficits. Granger causality analysis further indicated effective information flow from these FCs. This study showed significant correlation of cognitive impairment with aberrant FC in brain networks and hematological parameters in patients with β-TM. These results should advance our understanding of the neural mechanism underlying β-TM-related cognitive dysfunction and may serve as potential neuroimaging biomarkers for cognitive functioning in this population.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00277-026-06893-6.

Keywords: β-thalassemia major, Resting-state functional magnetic resonance imaging, Functional connectivity, Cognitive impairment, Granger causality analysis

Introduction

Thalassemia (TM) is a hereditary hemolytic disorder characterized by deficient hemoglobin synthesis and resultant chronic anemia [1, 2]. Although it is distributed worldwide [3], its prevalence is particularly high in certain regions, such as southern China. Among its subtypes, β-thalassemia major (β-TM) represents the most common clinical form, in which affected individuals typically develop profound anemia shortly after birth [4, 5]. Beyond hemoglobin deficiency, accumulating evidence indicates that β-TM major may exert broader deleterious effects on patients’ health [6, 7].

A landmark report in 2019 highlighted that individuals with β-TM frequently exhibit measurable cognitive impairments, underscoring the necessity of incorporating systematic neurocognitive assessment and psychological support into routine care [8]. At present, cognitive function is predominantly evaluated using standardized scales, such as the Montreal Cognitive Assessment (MoCA) and the Mini-Mental State Examination (MMSE) [9–11]. Besides, the modified MMSE (MMMSE), which requires less administration time while retaining the ability to assess key cognitive domains, has been particularly valuable in children and adolescents [9, 12]. Nevertheless, single-scale approaches are limited by their susceptibility to cultural and educational confounders.

Resting-state functional magnetic resonance imaging (rs-fMRI), which measures spontaneous fluctuations in blood oxygen level–dependent (BOLD) signals, provides a powerful tool for investigating the neural mechanisms underlying cognitive impairment in β-TM [13]. A core analytical framework in rs-fMRI is functional connectivity (FC), defined as the correlation of neural activity between spatially distinct brain regions [14].Previous rs-fMRI studies have reported widespread reductions in FC, involving the cerebellum (posterior lobules), Heschl’s gyrus, superior temporal gyrus, and motor–frontoparietal networks [15].

However, conventional FC methods primarily capture correlations between brain regions and often neglect potential temporal hierarchies among them [16]. Considering the systemic hematological abnormalities associated with β-TM, it is plausible that some brain regions play a dominant role in driving dysfunction across wider neural circuits. Understanding such directional influences is critical for elucidating disease-related mechanisms.

Granger causality analysis (GCA) offers a complementary approach that enables the investigation of directional and causal relationships within brain networks [17]. By assessing whether past activity in one region predicts current or future activity in another, GCA provides insight into effective connectivity and the dynamic flow of information within functional networks [18–20]. Integrating ROI-based FC with GCA may therefore offer complementary perspectives—correlational versus causal, static versus dynamic—yielding more robust neuroimaging evidence for cognitive dysfunction in β-TM. Nonetheless, studies employing such approaches in this population remain scarce.

In the present study, we prospectively enrolled a cohort of patients with β-TM major alongside age- and sex-matched healthy controls, all of whom underwent rs-fMRI scanning and standardized cognitive assessments. We hypothesized that β-TM patients would exhibit disrupted FC within and between large-scale brain networks supporting cognitive function, that these connectivity alterations would be associated with impairments across specific cognitive domains, and that certain networks would exert directional influences on others, contributing to cognitive dysfunction. To test these hypotheses, we adopted a stepwise analytical framework. First, ROI-to-ROI FC analysis was performed to identify network-level connectivity alterations between groups. Second, connectivity pairs showing significant group differences were examined for their associations with cognitive performance to establish behavioral relevance. Finally, GCA was applied to these cognition-related connectivity alterations to characterize the directionality of information flow among key networks, thereby extending correlational FC findings to a mechanistic, systems-level understanding of cognitive impairment in β-TM major.

Methods

Participants

From August 2021 to March 2024, consecutive patients with β-TM, aged 5–15 years, were prospectively recruited at the First Affiliated Hospital of Guangxi Medical University, P. R. China. Age- and sex-matched HCs were concurrently recruited from the local community. Diagnosis of β-TM was established based on clinical presentation, hematological indices, and molecular confirmation. All participants underwent brain MRI and comprehensive cognitive assessment using the Montreal Cognitive Assessment (MoCA) and Modified Mini-Mental State Examination (MMMSE). The inclusion and exclusion criteria for the two groups are presented in detail in Fig. 1. The final analytic cohort comprised 70 patients with β-TM and 64 HCs. Demographic characteristics and cognitive outcomes are summarized in Table 1, while hematological profiles (including complete blood counts, liver function tests, and serum ferritin) are provided in the supplementary file. The study protocol received approval from the local ethics committee, and written informed consent was obtained from all participants’ legal guardians.

Fig. 1.

Fig. 1

Open in a new tab

Flow chart illustrating the enrollment process of patients with β-thalassemia major (β-TM) and the healthy control (HC) participants

Table 1.

Demographic and clinical characteristics of the patients with β-TM and the HC participants in this study

Gender (male/female) β-TM (n = 70) HC (n = 64) p-value
39/31 42/22 0.244
Age 10.89 ± 2.24 10.19 ± 2.10 0.066
MoCA 22.07 ± 6.09 26.17 ± 2.78 <0.001
  Visual space and executive 2.99 ± 1.40 4.27 ± 0.95 <0.001
  Nominate 2.83 ± 0.45 2.89 ± 0.36 0.383
  Attention 4.79 ± 1.60 5.34 ± 0.95 0.016
  Verbal 1.74 ± 0.97 2.36 ± 0.78 <0.001
  Abstract 0.69 ± 0.67 1.05 ± 0.63 0.002
  Delay memory 3.43 ± 1.62 4.02 ± 1.24 0.021
  Directional 4.64 ± 1.48 5.25 ± 0.82 0.004
MMMSE 24.60 ± 5.01 27.31 ± 2.51 <0.001
  Time directive 4.04 ± 1.18 4.28 ± 0.79 0.177
  Space directive 3.61 ± 1.24 4.30 ± 0.87 <0.001
  Memory 2.79 ± 0.61 2.98 ± 0.13 0.012
  Attention calculation 3.79 ± 1.82 4.55 ± 1.17 0.005
  Remember 2.29 ± 0.95 2.52 ± 0.87 0.148
  Language 8.07 ± 1.17 8.69 ± 0.53 <0.001
Open in a new tab

Note: data are represented as the mean ± standard deviation (SD)

Abbreviations: MoCA, Montreal Cognitive Assessment; MMMSE, Modified Mini-Mental State Examination

Data acquisition and processing

For each subject, rs-fMRI data were acquired using a 3.0-Tesla GE MR scanner (SIGNA Premier, GE Healthcare, USA). Main parameters for rs-fMRI included 240 volumes, a repetition time of 2000 milliseconds (ms), an echo time of 30 ms, an in-plane resolution of 3.5 mm, 42 slices, and a slice thickness of 3.5 mm with a 0.3 mm gap. Analyses of fMRI data were performed using CONN [21] (RRID: SCR_009550) release 22.v2407 [22] and SPM [23] (RRID: SCR_007037) release 12.7771. The description of the procedure is included in the supplementary file.

Based on the results of the FC analysis, we identified ROIs where significant changes in FC occurred. Subsequently, we performed a correlation analysis between these brain regions and cognitive assessment scores.

GCA calculation

We used the copula-based GCA to further analyze the causal inferences between regions with significantly different FC metrics Compared to the traditional Granger causality calculation, this method works better for fitted or time-varying volatility time series and for uncovering non-linear causal relationships [17]. It should be noted that the aim of the present GCA was not to establish biophysical causality, but rather to characterize the directionality of information flow among cognition-related networks at a systems level. In this context, copula-based GCA provides a complementary perspective to functional connectivity by extending correlational findings to effective network interactions. The description of the procedure is included in the supplementary file.

To test the statistical significance of the copula-based Granger causality, we employed resampling techniques to construct a baseline null-hypothesis distribution. This distribution disrupted the causal dependency of interest while maintaining the other statistical properties of the data. A local bootstrap method (NBootstrap = 5000) was used to generate surrogate data for the null distribution, as proposed by Paparoditis and Politis [24]. A multitude of permuted versions of the original data set were created, wherein the trial order was randomly rearranged for each time series. The permutation of the trial order disrupted the relevant causal structure while maintaining the other statistical properties of the data, resulting in Granger causality values that were attributable to chance. Subsequently, Granger causality was performed on the permuted data sets to generate a null distribution. In a manner analogous to the single-trial procedure, the rank statistic was applied to compute the bootstrap p-value. The null hypothesis was rejected if the p-value of the estimated Granger causality was smaller than the significant p-value, such as 0.05. Subsequently, the abnormal effective Granger causality was compared between the patient and control groups using a two-factor analysis of variance (direction × group).

Statistical analysis

Statistical analyses were conducted using IBM SPSS Statistics version 29.0.1.0 (IBM Corporation, Armonk, NY, USA). Independent two-sample t-tests and Mann-Whitney U tests were used to analyze continuous variables. Chi-square tests were used for categorical variables. Pearson correlation or Spearman rank correlation was used to analyze the correlation between changes in FC and hematological indices, as well as cognitive testing scores. Statistical significance was set at a two-tailed p < 0 0.05. All analyses were adjusted for age and gender as covariates.

Results

Cognitive scores and hematological parameters for the study cohort

The MoCA and MMMSE scores of patients with β-TM were significantly lower than those of the HC group (p < 0.001). Detailed information for the cognitive scores is presented in Table 1. The hematological parameters in patients with β-TM, including white blood cell count, red blood cell count, hemoglobin levels, and serum ferritin levels are included in the supplementary file (Table S1).

The correlation between hematological parameters (including serum ferritin, red blood cell, and hemoglobin) and cognitive function in patients with β-TM is presented in Table S2. Several hematological parameters, such as corpuscular hemoglobin, hematocrit, total bilirubin, direct bilirubin, aspartate aminotransferase, and alanine aminotransferase, demonstrated statistically significant correlations with cognitive performance scores (p < 0.05), with correlation coefficients (r) approximately 0.3 (Table S2). However, no significant statistical correlation was found between serum ferritin and cognitive test results (p > 0.05)0.3.2. Altered FC in patients with β-TM.

Altered FC in patients with β-TM

Compared with the HCs, the β-TM group exhibited significantly reduced FC involving the language network, including the inferior frontal gyrus and posterior superior temporal gyrus; the dorsal attention network, including the intraparietal sulcus and frontal eye fields; the salience network, including the supramarginal gyrus, rostral prefrontal cortex, and anterior cingulate cortex; the frontoparietal network, including the posterior parietal cortex and lateral prefrontal cortex; and the sensorimotor network, including lateral and superior regions. In contrast, enhanced FC was observed among regions within the default mode network, including the posterior cingulate cortex, the visual medial network, and cerebellar network (including both posterior and anterior networks), as illustrated in Fig. 2.

Fig. 2.

Fig. 2

Open in a new tab

Comparison of functional connectivity (FC) of the patients with β-thalassemia (β-TM) major and the healthy control (HC) participants (β-TM > HC). The ends connected by red and blue represent different region of interest (ROI) areas. Blue line: Weakened FC. Red line: Enhanced FC. Abbreviation: PCC: posterior cingulate cortex; IPS: intraparietal sulcus; SMG: supramarginal gyrus; pSTG: posterior superior temporal gyrus; FEF: frontal eye fields; RPFC: rostral prefrontal cortex; ACC: anterior cingulate cortex; PPC: posterior parietal cortex; IFG: inferior frontal gyrus; LPFC: lateral prefrontal cortex; l: left side; r: right side. (p < 0.001 connection-level threshold, p-FDR < 0.001 cluster-level threshold)

Correlation between FC and cognitive performance

Partial correlation analyses identified significant associations between functional connectivity (FC) and cognitive performance across multiple large-scale networks. Connectivity between the anterior cerebellar network and the medial visual network was positively correlated with spatial orientation (r = 0.247, p = 0.043), attention and calculation (r = 0.264, p = 0.029), and total MMMSE score (r = 0.263, p = 0.030). Within sensorimotor–attention interactions, connectivity between the superior sensorimotor region and the left intraparietal sulcus within the dorsal attention network was positively correlated with orientation (r = 0.297, p = 0.014), whereas interhemispheric connectivity between the bilateral lateral sensorimotor cortices was negatively correlated with naming performance (r = − 0.300, p = 0.013).

Cross-network associations involving the language, salience, frontoparietal, and dorsal attention networks were also observed. Connectivity between the left posterior superior temporal gyrus within the language network and the left supramarginal gyrus within the salience network was positively associated with naming (r = 0.240, p = 0.048) and verbal performance (r = 0.245, p = 0.044). Connectivity between the right posterior superior temporal gyrus and the right posterior parietal cortex was positively correlated with attention (r = 0.369, p = 0.002), verbal performance (r = 0.331, p = 0.006), delayed memory (r = 0.344, p = 0.004), orientation (r = 0.271, p = 0.025), and MoCA score (r = 0.407, p = 0.001). Interhemispheric connectivity within the frontoparietal network further correlated with verbal performance (r = 0.371, p = 0.002), while connectivity between the left posterior parietal cortex and the left intraparietal sulcus was associated with naming (r = 0.298, p = 0.013), orientation (r = 0.243, p = 0.046), and MoCA score (r = 0.278, p = 0.022).

Additional associations were observed within executive and language-related networks. Interhemispheric connectivity between the bilateral rostral prefrontal cortices within the salience network correlated positively with visuospatial executive function (r = 0.279, p = 0.021), attention (r = 0.303, p = 0.012), and MoCA score (r = 0.297, p = 0.014). Interhemispheric connectivity between the bilateral inferior frontal gyri within the language network was positively correlated with spatial orientation (r = 0.269, p = 0.027), whereas connectivity between the right lateral prefrontal cortex and the left inferior frontal gyrus was associated with visuospatial executive function (r = 0.254, p = 0.036) (Fig. 3).

Fig. 3.

Fig. 3

Open in a new tab

Correlation between significantly altered functional connectivity (FC) regions and cognitive subscale scores. The matrix displays the complete correlation pattern, with black frame highlighting correlations between regional FC values and cognitive subscale scores. Color intensity in the matrix corresponds to correlation coefficients and the asterisks denote statistically significant differences (*p < 0.05, **p < 0.01, ***p < 0.001). Abbreviation: MMMSE: Modified Mini-Mental State Examination; MoCA: Montreal Cognitive Assessment; VSEF: visual-spatial executive function; atten_calc: attention and calculation; PCC: posterior cingulate cortex; IPS: intraparietal sulcus; SMG: supramarginal gyrus; pSTG: posterior superior temporal gyrus; FEF: frontal eye fields; RPFC: rostral prefrontal cortex; ACC: anterior cingulate cortex; PPC: posterior parietal cortex; IFG: inferior frontal gyrus; LPFC: lateral prefrontal cortex;; l: left side; r: right side

At the network level, the observed connectivity alterations primarily involved the frontoparietal network (FPN), salience network (SN), dorsal attention network (DAN), language network, and cerebellar network. Reduced connectivity within and between the FPN, SN, DAN, and language networks was consistently associated with impairments in attention, executive function, memory, and language performance, highlighting these networks as core substrates of cognitive dysfunction in β-TM. In contrast, increased connectivity involving the cerebellum and medial visual and default mode networks (DMN) showed positive associations with cognitive scores.

Disrupted effective connectivity in patients with β-TM

GCA revealed the directional patterns of significant effective connectivity among brain networks in the β-TM group. Both the anterior cerebellum and posterior cerebellum exerted significant causal influences on the medial visual cortex. The posterior parietal cortex within the frontoparietal control network showed a significant directed causal effect on the posterior superior temporal gyrus of the language network.

Within the salience network, the supramarginal gyrus exerted a significant causal influence on the lateral region of the sensorimotor network, while the right rostral prefrontal cortex (frontal pole) within the salience network showed a significant causal drive toward the anterior cingulate cortex of the same network; moreover, a bidirectional causal interaction was observed between the bilateral rostral prefrontal cortices within the salience network.

Within the sensorimotor network, the superior sensorimotor region exerted a significant causal influence on the intraparietal sulcus of the dorsal attention network, and the right lateral sensorimotor region showed a significant directed causal effect on the left lateral sensorimotor region. All Granger causality results survived bootstrap significance testing (α = 0.05, number of bootstrap iterations = 5000) (Fig. 4).

Fig. 4.

Fig. 4

Open in a new tab

Schematic diagram of functional connectivity among different brain regions based on Granger causal interactions in patients with β-thalassemia major (β-TM). The direction of the arrows indicates the direction of information flow in a causal relationship. Abbreviation: PCC: posterior cingulate cortex; IPS: intraparietal sulcus; SMG: supramarginal gyrus; pSTG: posterior superior temporal gyrus; FEF: frontal eye fields; RPFC: rostral prefrontal cortex; ACC: anterior cingulate cortex; PPC: posterior parietal cortex; IFG: inferior frontal gyrus; LPFC: lateral prefrontal cortex; l: left side; r: right side

Discussion

This study demonstrates a complex reorganization of brain networks in children and adolescents with β-TM, characterized by both disrupted FC and compensatory adaptations. We identified specific FC alterations linked to cognitive domains, including memory and attention, and GCA revealed directional information flow within key networks.

Specifically, with respect to cognitive performance, patients with β-TM demonstrated significantly lower MoCA and MMMSE scores compared to healthy controls, indicating potential impairments across multiple domains, including attention, calculation, orientation, and language. These findings are consistent with previous reports [6, 7]. We next examined the correlation between hematological parameters and cognitive scores. Although several indices—such as mean corpuscular hemoglobin, hematocrit, total bilirubin, direct bilirubin, aspartate aminotransferase, and alanine aminotransferase—showed statistically significant associations with cognitive measures (p < 0.05), the correlations were weak, and the modest effect sizes limit their clinical relevance [25]. The lack of a significant association between serum ferritin and cognitive performance likely reflects the limited ability of peripheral blood markers to capture cerebral iron status, given the tight regulation of iron transport across the blood-brain barrier [26–28] and the influence of clinical factors such as inflammation and chelation therapy. This discrepancy underscores the value of neuroimaging-based biomarkers, including quantitative susceptibility mapping and other iron-sensitive MRI techniques, which can directly quantify regional brain iron deposition and may provide a more accurate link between iron overload and cognitive dysfunction in β-TM [29].

At the intra-network level, FC between interhemispheric regions of higher-order cognitive networks, including salience, frontoparietal, language, and dorsal attention networks was significantly attenuated. Impaired interhemispheric communication may underlie observed deficits in executive control and attention [30]. Specifically, reduced connectivity between the dorsolateral prefrontal cortex and posterior parietal cortex correlated with visuospatial and executive function, supporting the role of the frontoparietal network as a flexible hub for cognitive control [31]. Similarly, attenuated connectivity of the rostral prefrontal cortex, a core salience network node, was associated with poorer attention and executive function, consistent with prior observations in β-TM [15].

Inter-network analysis revealed diminished FC between the language network (posterior superior temporal gyrus) and the frontoparietal network (posterior parietal cortex), as well as between frontoparietal and dorsal attention networks. Notably, Connectivity between the right posterior parietal cortex and right posterior superior temporal gyrus correlated positively with attention, language, delayed memory, and orientation scores. This finding aligns with the understanding that delayed memory retrieval requires top-down modulation from control networks like the frontoparietal network [31–33]. Our GCA results further supported this, indicating a unidirectional causal influence from the right posterior parietal cortex to the right posterior superior temporal gyrus, reflecting top-down attentional modulation of language processing regions [34]. Besides, Impaired connectivity along this pathway may reflect deficient executive control and resource allocation, analogous to disrupted frontoparietal connectivity observed in other conditions such as traumatic brain injury [35].

Notably, some potential evidence of compensatory reorganization was observed. Although connectivity between the supramarginal gyrus (a salience-network node) and both language and sensorimotor networks was reduced, GCA revealed a directed influence from the supramarginal gyrus to lateral sensorimotor regions, which may reflect recruitment of auxiliary processing resources to support other functions [36]. The strength of this cross-hemispheric connection correlated positively with spatial orientation, attention, and language performance in our cohort, suggesting that when primary pathways are compromised, auxiliary interregional links are engaged to preserve function. This interpretation is biologically plausible: animal models show that iron overload produces hippocampal injury associated with spatial-memory impairments, providing a potential pathological substrate for spatial disorientation in β-TM and a rationale for the emergence of compensatory circuits [37], [38].

Further evidence of compensation was observed in the cerebellum, a structure that is increasingly recognized for its role in higher-order cognition [39, 40]. Specifically, the cerebellum exhibited heightened connectivity with cortical regions, predominantly involving the medial visual network. Complementing this observation, our GCA analysis revealed a unidirectional causal influence from the cerebellum to the medial visual network. In the context of impaired top-down control from the frontoparietal and dorsal attention networks, such cerebellar-driven modulation may enhance processing efficiency or stabilize visual input, thereby compensating for deficient cortical signaling [41]. This interpretation is further supported by the positive correlations observed between cerebellar–visual connectivity and performance in attention, calculation, and spatial orientation, aligning with prior evidence of cerebellar contributions to cognitive processes in healthy populations [42, 43]. Collectively, our findings of widespread cortical hypoconnectivity alongside targeted cerebellar-cortical hyperconnectivity, support a dynamic, biphasic neuropathological model for β-TM: an initial recruitment of compensatory mechanisms, such as those involving the cerebellum, is challenged by the persistent and progressive burden of chronic disease [44, 45].

In summary, this study demonstrates that patients with β-TM exhibit altered connectivity within key cognitive networks, and the extent of these alterations correlates with their degree of cognitive impairment. We also identified enhanced, directional cerebellar-cortical connectivity, suggesting a potential compensatory mechanism to preserve cognitive function. These findings offer novel insights into the complex interplay between systemic disease and brain function in β-TM. The specific patterns of network dysfunction and compensation identified here may serve as potential neuroimaging biomarkers for monitoring disease progression and the associated risk of cognitive decline.

Limitations

This study has several limitations. First, the modest sample size limits our ability to examine potential confounding factors, such as disease duration, transfusion frequency, or specific chelation regimens. Second, the application of GCA to fMRI data should be interpreted with caution, as the low temporal resolution of the fMRI signal and regional variability in hemodynamic responses may confound causal inferences. Finally, reliance on a single neuroimaging modality (rs-fMRI) and static functional connectivity analyses provides an incomplete view of the neural mechanisms involved. Future studies should adopt longitudinal, multi-modal approaches, incorporating techniques such as diffusion tensor imaging for white matter integrity [46], quantitative susceptibility mapping for brain iron deposition [47], and arterial spin labeling for cerebral perfusion [48], to achieve a more comprehensive understanding of the structural and functional brain alterations underlying cognitive outcomes in β-TM.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1 (DOCX 26.1 KB) (26.2KB, docx)

Acknowledgements

Not applicable.

Author contributions

Haifeng Zheng and Hang Yu contributed equally to this work as co-first authors. Haifeng Zheng’s primary contributions included conceptualization, data curation, formal analysis, writing the original draft, review and editing. Hang Yu’s primary contributions encompassed methodology, formal analysis, validation, visualization, and paper review and editing). Xingye Yang, Xi Deng and Meiru Bu’s main contributions are for data curation. Yuhong Qin and Bihong T. Chen were involved in paper writing (review and editing). Yifeng Wang and Muliang Jiang, as co-corresponding authors, oversaw the study, providing supervision and paper writing (review and editing). Additionally, Muliang Jiang provided support in conceptualization, resources, and validation. All authors reviewed the final version of the manuscript. All authors read and approved the final version of this manuscript.

Funding

This work was supported by National Natural Science Foundation of China (Grant No. 82260344) and the Natural Science Foundation of Guangxi (Grant No. GUIKE AB25069012).

Data availability

The datasets used and/or analyzed in the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study received approval from the First Affiliated Hospital of Guangxi Medical University Ethical Review Committee (IRB No. 2023-S615-01)

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Haifeng Zheng, Hang Yu and Xingye Yang contributed equally to this work.

Contributor Information

Yifeng Wang, Email: wyf@sicnu.edu.cn.

Muliang Jiang, Email: jmlgxmu@gmail.com.

References

  • 1.Bajwa H, Basit H (2025) Thalassemia In: StatPearls. StatPearls Publishing Copyright © 2025, StatPearls Publishing LLC., Treasure Island (FL)
  • 2.Taher AT, Longo DL, Musallam KM, Cappellini MD (2021) β-Thalassemias. N Engl J Med 384:727–743. 10.1056/NEJMra2021838 [DOI] [PubMed] [Google Scholar]
  • 3.Tuo Y, Li Y, Li Y, Ma J, Yang X, Wu S et al (2024) Global, regional, and National burden of thalassemia, 1990–2021: a systematic analysis for the global burden of disease study 2021. eClinicalMedicine 72:101016jeclinm2024102619 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.De Simone G, Quattrocchi A, Mancini B, di Masi A, Nervi C, Ascenzi P (2022) Thalassemias: from gene to therapy. Mol Aspects Med 84:101028. 10.1016/j.mam.2021.101028 [DOI] [PubMed] [Google Scholar]
  • 5.Brusgaard K, Yin A, Li B, Luo M, Xu L, Wu L et al (2014) The prevalence and molecular spectrum of α- and β-Globin gene mutations in 14,332 families of Guangdong Province, China. PLoS ONE. 9. 10.1371/journal.pone.0089855 [DOI] [PMC free article] [PubMed]
  • 6.Bizri M, Koleilat R, Akiki N, Dergham R, Mihailescu AM, Bou-Fakhredin R et al (2024) Quality of life, mood disorders, and cognitive impairment in adults with β-thalassemia. Blood Rev 65. 10.1016/j.blre.2024.101181 [DOI] [PubMed]
  • 7.Nevruz O, Ulas U, Cetin T, Kutukcu Y, Kurekci A (2006) Cognitive dysfunction in β-thalassemia minor. Am J Hematol 82:203–207. 10.1002/ajh.20798 [DOI] [PubMed] [Google Scholar]
  • 8.Rund D (2019) Cognition in thalassaemia: the next milestone. Br J Haematol 186:511–512. 10.1111/bjh.15956 [DOI] [PubMed] [Google Scholar]
  • 9.Ouvrier RA, Goldsmith RF, Ouvrier S, Williams IC (1993) The value of the Mini-Mental state examination in childhood: A preliminary study. J Child Neurol 8:145–148. 10.1177/088307389300800206 [DOI] [PubMed] [Google Scholar]
  • 10.Pinto TCC, Machado L, Bulgacov TM, Rodrigues-Júnior AL, Costa MLG, Ximenes RCC, Sougey EB (2019) Is the Montreal cognitive assessment (MoCA) screening superior to the Mini-Mental state examination (MMSE) in the detection of mild cognitive impairment (MCI) and alzheimer’s disease (AD) in the elderly? Int Psychogeriatr 31:491–504. 10.1017/s1041610218001370 [DOI] [PubMed] [Google Scholar]
  • 11.Islam N, Hashem R, Gad M, Brown A, Levis B, Renoux C et al (2023) Accuracy of the Montreal cognitive assessment tool for detecting mild cognitive impairment: A systematic review and meta-analysis. Alzheimer’s Dement 19:3235–3243. 10.1002/alz.13040 [DOI] [PubMed] [Google Scholar]
  • 12.Ibrahim BA, Hussein SA, Abdullah WH (2023) COGNITIVE FUNCTIONS IN CHILDREN WITH TYPE I DIABETES. Wiad Lek 76:944–950. 10.36740/WLek202305108 [DOI] [PubMed] [Google Scholar]
  • 13.Amemiya S, Takao H, Abe O (2024) Resting-State fMRI: emerging concepts for future clinical application. J Magn Reson Imaging 59:1135–1148. 10.1002/jmri.28894 [DOI] [PubMed] [Google Scholar]
  • 14.de Thiebaut M, Forkel SJ (2022) The emergent properties of the connected brain. Science 378:505–510. 10.1126/science.abq2591 [DOI] [PubMed] [Google Scholar]
  • 15.Xu S, Wu J, Li Y, Liu X, Liu S, Wang X et al (2024) Brain functional connectivity as a mediator between hematological metrics and cognitive decline in children with beta-thalassemia major. BioRxiv 202420112006622267. 10.1101/2024.11.06.622267 [DOI] [PMC free article] [PubMed]
  • 16.Schwab S, Harbord R, Zerbi V, Elliott L, Afyouni S, Smith JQ et al (2018) Directed functional connectivity using dynamic graphical models. NeuroImage 175:340–353. 10.1016/j.neuroimage.2018.03.074 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Hu M, Liang H (2014) A copula approach to assessing Granger causality. NeuroImage 100:125–134. 10.1016/j.neuroimage.2014.06.013 [DOI] [PubMed] [Google Scholar]
  • 18.Yang C, Xiao K, Ao Y, Cui Q, Jing X, Wang Y (2023) The thalamus is the causal hub of intervention in patients with major depressive disorder: evidence from the Granger causality analysis. Neuroimage Clin 37:103295. 10.1016/j.nicl.2022.103295 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Wang Y, Yang C, Li G, Ao Y, Jiang M, Cui Q et al (2023) Frequency-dependent effective connections between local signals and the global brain signal during resting-state. Cogn Neurodyn 17:555–560. 10.1007/s11571-022-09831-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Friston K, Moran R, Seth AK (2013) Analysing connectivity with Granger causality and dynamic causal modelling. Curr Opin Neurobiol 23:172–178. 10.1016/j.conb.2012.11.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Whitfield-Gabrieli S, Nieto-Castanon A (2012) Conn: A functional connectivity toolbox for correlated and anticorrelated brain networks. Brain Connect 2:125–141. 10.1089/brain.2012.0073 [DOI] [PubMed] [Google Scholar]
  • 22.Nieto-Castanon A, Whitfield-Gabrieli S (2022) CONN functional connectivity toolbox: RRID SCR_009550, release 22 [DOI] [PubMed]
  • 23.Penny WD, Friston KJ, Ashburner JT, Kiebel SJ, Nichols TE (2011) Statistical parametric mapping: the analysis of functional brain images. Elsevier
  • 24.Paparoditis E, Politis DN (2003) Residual-based block bootstrap for unit root testing. Econometrica 71:813–855 [Google Scholar]
  • 25.Schober P, Boer C, Schwarte LA (2018) Correlation coefficients: appropriate use and interpretation. Anesth Analg 126:1763–1768. 10.1213/ane.0000000000002864 [DOI] [PubMed] [Google Scholar]
  • 26.Origa R, Galanello R, Ganz T, Giagu N, Maccioni L, Faa G, Nemeth E (2007) Liver iron concentrations and urinary Hepcidin in beta-thalassemia. Haematologica 92:583–588. 10.3324/haematol.10842 [DOI] [PubMed] [Google Scholar]
  • 27.Argyropoulou MI, Kiortsis DN, Astrakas L, Metafratzi Z, Chalissos N, Efremidis SC (2007) Liver, bone marrow, pancreas and pituitary gland iron overload in young and adult thalassemic patients: a T2 relaxometry study. Eur Radiol 17:3025–3030. 10.1007/s00330-007-0683-1 [DOI] [PubMed] [Google Scholar]
  • 28.Noetzli LJ, Carson SM, Nord AS, Coates TD, Wood JC (2008) Longitudinal analysis of heart and liver iron in thalassemia major. Blood 112:2973–2978. 10.1182/blood-2008-04-148767 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Bayraktaroglu S, Karadas N, Onen S, Karapinar DY, Aydinok Y (2022) Modern management of iron overload in thalassemia major patients guided by MRI techniques: real-world data from a long-term cohort study. Ann Hematol 101:521–529. 10.1007/s00277-021-04748-w [DOI] [PubMed] [Google Scholar]
  • 30.Jin X, Liang X, Gong G (2020) Functional integration between the two brain hemispheres: evidence from the homotopic functional connectivity under resting state. Front Neurosci 14:932. 10.3389/fnins.2020.00932 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Marek S, Dosenbach NUF (2018) The frontoparietal network: function, electrophysiology, and importance of individual precision mapping. Dialogues Clin Neurosci 20:133–140. 10.31887/DCNS.2018.20.2/smarek [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Westphal AJ, Wang S, Rissman J (2017) Episodic memory retrieval benefits from a less modular brain network organization. J Neurosci 37:3523–3531. 10.1523/jneurosci.2509-16.2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Sestieri C, Corbetta M, Romani GL, Shulman GL (2011) Episodic memory retrieval, parietal cortex, and the default mode network: functional and topographic analyses. J Neurosci 31:4407–4420. 10.1523/jneurosci.3335-10.2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Lindell AK (2006) In your right mind: right hemisphere contributions to Language processing and production. Neuropsychol Rev 16:131–148. 10.1007/s11065-006-9011-9 [DOI] [PubMed] [Google Scholar]
  • 35.Fortenbaugh FC, DeGutis J, Esterman M (2017) Recent theoretical, neural, and clinical advances in sustained attention research. Ann N Y Acad Sci 1396:70–91. 10.1111/nyas.13318 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Wandelt SK, Bjånes DA, Pejsa K, Lee B, Liu C, Andersen RA (2024) Representation of internal speech by single neurons in human supramarginal gyrus. Nat Hum Behav 8:1136–1149. 10.1038/s41562-024-01867-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Sun Y, Tian B, Liang J, Bu M, Deng X, Deng K et al (2024) Hippocampal iron overload and Spatial reference memory impairment: insights from a rat model. Neurosci Lett 843:138014. 10.1016/j.neulet.2024.138014 [DOI] [PubMed] [Google Scholar]
  • 38.Deng X, Bu M, Liang J, Sun Y, Li L, Zheng H et al (2025) Relationship between cognitive impairment and hippocampal iron overload: A quantitative susceptibility mapping study of a rat model. NeuroImage 306:121006. 10.1016/j.neuroimage.2025.121006 [DOI] [PubMed] [Google Scholar]
  • 39.Koziol LF, Budding D, Andreasen N, D’Arrigo S, Bulgheroni S, Imamizu H et al (2014) Consensus paper: the cerebellum’s role in movement and cognition. Cerebellum 13:151–177. 10.1007/s12311-013-0511-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Abdallah M, Zahr NM, Saranathan M, Honnorat N, Farrugia N, Pfefferbaum A et al (2021) Altered Cerebro-Cerebellar dynamic functional connectivity in alcohol use disorder: a Resting-State fMRI study. Cerebellum 20:823–835. 10.1007/s12311-021-01241-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Nguyen KP, Person AL (2025) Cerebellar circuit computations for predictive motor control. Nat Rev Neurosci. 10.1038/s41583-025-00936-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Brissenden JA, Somers DC (2019) Cortico-cerebellar networks for visual attention and working memory. Curr Opin Psychol 29:239–247. 10.1016/j.copsyc.2019.05.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Kellermann T, Regenbogen C, De Vos M, Mößnang C, Finkelmeyer A, Habel U (2012) Effective connectivity of the human cerebellum during visual attention. J Neurosci 32:11453–11460. 10.1523/jneurosci.0678-12.2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Jehee JF, Ballard DH (2009) Predictive feedback can account for biphasic responses in the lateral geniculate nucleus. PLoS Comput Biol 5:e1000373. 10.1371/journal.pcbi.1000373 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Carter CS (2025) The psychedelic-peptide paradox: a hormetic hypothesis. Compr Psychoneuroendocrinol 23:100303. 10.1016/j.cpnec.2025.100303 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Yen C, Lin CL, Chiang MC (2023) Exploring the frontiers of neuroimaging: A review of recent advances in Understanding brain functioning and disorders. Life (Basel) 13. 10.3390/life13071472 [DOI] [PMC free article] [PubMed]
  • 47.Harada T, Kudo K, Fujima N, Yoshikawa M, Ikebe Y, Sato R et al (2022) Quantitative susceptibility mapping: basic methods and clinical applications. Radiographics 42:1161–1176. 10.1148/rg.210054 [DOI] [PubMed] [Google Scholar]
  • 48.Jaafar N, Alsop DC (2024) Arterial spin labeling: key concepts and progress towards use as a clinical tool. Magn Reson Med Sci 23:352–366. 10.2463/mrms.rev.2024-0013 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1 (DOCX 26.1 KB) (26.2KB, docx)

Data Availability Statement

The datasets used and/or analyzed in the current study are available from the corresponding author upon reasonable request.

Articles from Annals of Hematology are provided here courtesy of Springer

RESOURCES