This study demonstrated extensive brain white matter alterations in pediatric SMA types 2 and 3, which is responsive to the effects of SMN protein deficit on brain white matter since individuals with SMA types 2 and 3 are less likely to be affected by ischemic-hypoxic encephalopathy [13]. There are several important findings in this study. Firstly, we found widespread white-matter changes among many fiber groups of the brain, including projecting, commissural, association, and limbic system fibers. Secondly, there is a reduction in the ability to integrate and the efficiency of information transit across the entire brain network in SMA patients. Finally, at the nodal level, altered nodal profiles were mainly located in right supramarginal gyrus, right orbital part of superior frontal gyrus, right supplementary motor area, and left median cingulate and paracingulate gyri, which are known to be involved in cognitive function, emotion regulation, motor planning and execution, and verbal working memory [25,26,27,28,29].

In SMA patients, widespread white matter alterations were shown in our study, with the parameters FA and AD of the SMA group being smaller than those of the HC group, and RD of the SMA group being larger than those of the HC group. A decrease in the FA value suggests a reduction in integrity and damage to the white matter structure, from gliosis, etc., axonal injury, and demyelination are among the potential causes. Also, diffusion along the axon direction decreases with a decreasing AD value, indicating axonal injury or the breakdown of fiber bundle consistency. Additionally, an increase in RD illustrates a loss of myelin or compromised integrity [14]. We propose that a key contributor to the cognitive impairments and impaired motor function linked to SMA is aberrant white matter fiber development. The corpus callosum has the function of transmitting information across the midline of the brain, which is associated with cognition, memory, and emotional modulation [30]. The left SLF-I is linked to verbal working memory, with SLF-III supporting the working memory phonological loop. Right hemisphere SLF-I/II damage impairs visuospatial working memory [31]. The functions of the corticospinal tract are the modulation of afferent signals, spinal reflexes, and motor neuron activation. The corticospinal tract holds a crucial position within the motor system, as it facilitates voluntary movements in the distal extremities [32]. We thought that abnormalities in the development of the corpus callosum, superior longitudinal fasciculus, and corticospinal tract might be associated with SMA-related cognitive function and motor function changes. A recent study showed that patients with type 2 and 3 SMA had significantly lower bilateral white matter volumes in the frontal and parietal lobes [33], which is consistent with our findings.

Besides, deterioration of clinical symptoms and muscle atrophy evaluated by HFMSE scores were found to be accompanied by decreased AD values in specific regions. The areas included right anterior thalamic radiation, corticospinal tract and superior longitudinal fasciculus, which are related to motor function [32] and executive function [31]. Changes in the brain white matter may be associated with decreased motor function due to the disease.

It is well-known that information transmission is the primary role of brain white matter. The aberrant development of white matter fibers usually causes abnormal white matter network integration, which is also seen in SMA patients. The SMA group had higher Lp, γ, and σ and lower Eglob than the HC group. Lower Eglob and higher Lp indicate a decline in the ability for integration and the efficiency of information transport in parallel across the entire brain network [34]. It suggests that the SMA white matter network had a less optimal topological organization and a tendency to become high-cost and low-profit that more unlike a brain’s small world network [15, 35]. Therefore, SMA may cause dysfunctions of large-scale spatially distributed neural networks, particularly whole-brain integration deficits. This finding aids in the comprehension of SMA symptoms.

In a further analysis of node properties, multiple node properties were found to be different in the anatomical networks of the SMA group compared to the controls. Those brain regions were associated with alterations in cognitive function, emotion regulation, motor planning and execution, and verbal working memory [25,26,27,28,29]. For the cognitive function, several previous studies have identified varying degrees of cognitive dysfunction in patients with SMA [16, 36,37,38,39]. Disruption of the white matter microstructure of the brain has been reported to alter the cognitive network of the brain and is one of the key pathological underpinnings of cognitive impairment [40]. Further more in-depth studies following up SMA patients will be crucial to explain this finding. Previous study also has found that topological properties in the cortical-limbic-cerebellum circuit was disrupted in SMA [41]. In our study, the brain regions where node parameters have changed are widespread. We have not yet found a common neural pathway for the above different brain regions. They are involved in multiple neural networks, including the DMN, though their involvement varies. Overall, they support various cognitive, emotional, and motor functions. In the future, we hope to clarify the neural pathways that cause SMA patients’ diseases through further research.

Altered nodal metrics in SMA were found in right supplementary motor area, which is usually regarded as an essential region for motor planning and execution [25]. Abnormal alterations in this area suggested that brain involvement may be associated with symptoms of SMA, especially those of impaired motor function. In the node parameters, the abnormality of the supplementary motor area and the HFMSE scores are not correlated, which may not represent that the area is not correlated in the real situation, but only due to the limitation of the relatively indirect measurement method of network analysis. In the future, we will adopt more advanced methods for further research. With the advent of pharmacotherapy, motor function will be enhanced in SMA patients. Thus, further longitudinal studies are essential to determine whether connectivity in this region will improve after treatment.

The possible causes of abnormal white matter fibers in SMA patients are complex. Firstly, low-level SMN protein caused by genetic mutations is believed to have a significant impact on the brain. Generally, all major white matter tracts are present by the end of normal gestation (37–42 weeks). Processes including myelination and synaptogenesis occur rapidly during the first 2–3 years of life, and ongoing brain remodeling continues into early adulthood [42]. The observed alteration in white matter is probably the result of a delated myelin maturation throughout development.

Also, the impact of limb disuse on the brain cannot be ignored. In the other functional neurological disorders, subtle changes in brain area volume and cortical thickness were also found and were considered to be secondary changes due to limb disuse in the case of thalamic or motor areas [43]. Similarly, the limb disuse due to muscle weakness in SMA patients should also be considered as an influencing factor of white matter changes.

Besides, the age-related changes may have a small impact on the results. But it has been minimized to a relatively low level by matching the age and gender of the participants and enrolling participants aged 5 to 18 who were in a relatively stable stage of brain development.

This study also has a few limitations. First, this was a single-center study with a small number of participants and a limited sample size due to the low morbidity of SMA. Second, the study used a cross-sectional design. A longitudinal design is a natural extension of our proof-of-concept findings. Third, due to failure to collect scales for patient cognitive assessment, we did not capture correlations between neuroimaging measures and cognitive assessments. In the future, multi-center research with large sample sizes and even longitudinal studies combining structural and functional MRI may allow for a deeper comprehension of the neurobiological aspects of SMA.