Aneurysmal subarachnoid hemorrhage (aSAH) is one common disease which causes vascular dementia. It is traditionally regarded as one severe incident with high mortality and morbidity (Lovelock et al., 2010). The hemorrhagic strike is sudden and deadly because of a high prevalence of occult or undetected intracranial aneurysms (Vernooij, 2007; Vlak et al., 2011). But in the recent decades, there is a tremendous improvement on the overall survival rate of aSAH patients due to the great progress made on early diagnosis and timely intervention.

As the mortality rate is dropping in the last two decades, more patients experience severe neurological sequela after survival. Several studies reported that even those survivors categorized into “mild disability” also suffered from various cognitive impairments such as working memory deficit, aphasia or alexia (Ørbo et al., 2008). That increases the number of patients who needs rehabilitative care and return to their normal life. As a result, it is of great value to study the cognitive impairment after aSAH. New strategy or improvement of intervention and rehabilitation needs theoretic foundation, which comes from specific description of neural behavior and recovery process.

According to the previous research, short-term memory has been listed as the most important sequela (Mayer et al., 2002; Anderson et al., 2006; Al-Khindi et al., 2010; Scott et al., 2010). Particularly, reports pointed out that working memory deficit is often affected (Sheldon et al., 2012, 2013). This persistent cognitive problem lays a great impact on patients' daily life. Although a number of rehabilitative methods have been applied, most approaches are empirical or experimental (Wong et al., 2016; Huenges Wajer et al., 2017; Stabel et al., 2017). Lacking theoretical support somehow limits the improvement of current rehabilitation. Of course, previous studies have provided insight into the functional alterations of the working memory deficit after aSAH. It reported that aSAH patients with cognitive impairment displayed significant decline of spontaneous neural activity in several cerebral regions (Su, 2018). However, there are still questions about the specific cortical changes of aSAH patients with working memory deficit. (1)To what extent is the brain network of those with memory deficit affected, whole-brain network level or sub-network level? (2) How is the working memory network organized when isolated regions have been proved dysfunctional? Although the existing evidence revealed some functional changes, identifying worthwhile large-scale changes would be a vital advance for clarifying the underlying mechanism of working memory deficit.

In this study, we would extract several sub-networks including working memory network from the resting-state fMRI data. The topological changes of networks instead of activation status of one brain region would be analyzed. We aim to provide novel information about the characteristics of brain organization, especially “small-world” property (Sporns and Zwi, 2004). The comparison would be performed in the aSAH patients with and without working memory deficit in order to reveal more information about the associated neural plasticity.

We compare the gender as well as the length of hospitalization and the statistical result shows no significant difference. The AVLT scaling result is 9.33 ± 0.93/12.45 ± 1.02(Mean ± SEM) in control group and 6.21 ± 0.76/9.58 ± 0.87 (Mean ± SEM) in working memory deficit group. The statistical analysis confirms significant decrease of delayed memory function in working memory deficit group (Table 1).

After ICA procedure, we extract 40 components and establish totally eight sub-networks from the resting data including working memory network, basal ganglia network, dorsal default mode network, visual network, left executive control network, sensorimotor network, visual spatial network (Figure 1). In the analysis of small-worldness analysis, aSAH group with working memory deficit shows no significant difference of clustering coefficient with control group (Figure 2). But it shows significant decrease of characteristic path length than control group (Figure 3). As for the analysis of local and global efficiency, the result shows no significant difference between groups (Figure 4).

The detailed alteration of cognitive impairment after aSAH has been always a vital problem for researchers (Welmer et al., 2006; Al-Khindi et al., 2010; Ali et al., 2020). Because improvement on current treatment or development of new interventional strategy depends on our understanding of neural mechanism. However, there is still a long way before the thorough clarification of neural basis, which leads to multiple cognitive impairment including working memory deficit. In fact, the exact neurological change of working memory deficit after aSAH remains unclear.

Our study is motivated by previous research, which indicated persistent memory dysfunction after recovery of aSAH (Ørbo et al., 2008; Schweizer and Macdonald, 2010). Task-dependent BOLD experiment has been utilized to reveal the abnormal cortical activation which is potentially associated with memory deficit (Ellmore et al., 2013). Nevertheless, it is still not sufficient for a detailed description on the neurological change. Because isolated activation of certain brain regions fails to give enough attention to the connective situation between those regions. It is possible to reveal more details about working memory deficit if we consider more about the characteristic of information transmission from an angle of brain network.

On the other hand, the cognitive impairment is believed to trigger very limited response in whole brain scale compared with other task of strong activation. In order to increase our sensitivity of detection, we also include regional functional network to our analysis. The working memory network is localized and identified by ICA, which enables us to extract individual sub-network from group-level data.

Our study provides insight to the topological property of working memory deficit after aSAH. The most important findings could be concluded as followed, (1) The significant alteration of working memory deficit stays in a limited extent, which lies in regional working memory network rather than whole-brain network. (2) Working memory dysfunction promotes higher overall routing efficiency of regional functional network.

The most important result of this study is the significant decrease of characteristic path length. That indicates the surprising increase of overall routing efficiency of working memory network in aSAH patients with working memory deficit, which somehow fails to be in line with former studies (Chung et al., 2019). Previous studies demonstrated that working memory deficit was related with reduced connectivity within the network (Arciniega et al., 2021; Vannest et al., 2021)and other neuron loss or dysfunctionality (Fang et al., 2018; Cheng et al., 2019). Most reports focus on the direct hemodynamic response or adjacent connectivity of regions such as left dorsolateral prefrontal cortex (DLPFC) (Burgess et al., 2010) or parietal cortex (Olesen et al., 2004). Some existing evidence suggested subjects with working memory deficit requires more executive control, and therefore more brain activity, compared to healthy (or faster) individuals in order to perform successfully (Postle et al., 1999; Chang et al., 2011). An exact idea to explain this is the Compensation-Related Utilization of Neural Circuits Hypothesis (Reuter-Lorenz and Cappell, 2008), whereby patients with working memory deficit have to recruit more neuronal resources at lower loads than those with normal memory capacity leaving fewer resources for processing at higher loads. Similar result was found in patients of ADHD with working memory deficit. The imaging study discovered decreased efficiency of DLPFC for high-load visuospatial working memory and greater reliance on posterior spatial attention circuits to store and update spatial position than healthy control youth (Bédard et al., 2014). Actually, it was unclear at the outset whether memory dysfunction would be accompanied by reduced or increased activation. But we could say that the direct task-dependent imaging and large-scale network analysis provide two dimension of information about the detailed neural alteration contributing to the working memory deficit. Of course, the current theory needs larger sample size for confirmation in further study.

We also noticed the insignificant working network in left hemisphere. Previous studies (Owen et al., 2005) reported activation in the left inferior parietal lobe and left precunues increased along with increasing age. Those regions were implicated in WMN studies. According to this theory, the subjects recruited in our study would display progressive recruitment of these task related frontal and parietal regions that underpin the functional maturation of working memory. In this circumstances, the left working memory system showed more robust architecture than the right one in the face of disease strike.

We perform independent component analysis for separating noise and functional sub-network. We concatenate all the data from both groups and then perform group-level dimensional reduction. It is true that the included images from two groups might differ in their brain signal. But we reason that the extracted spatial pattern of components should be unified in both groups for incoming statistical comparison. As a result, the spatial maps of component representing working memory network is not far different from previous human studies (Damoiseaux et al., 2006, 2008). In addition, our result does not include another pattern of working memory network, in which most of the significant voxels are localized in the left hemisphere. We think this may attribute to the limited sample size of included subjects. The component of right hemispheric working memory network fails to achieve the threshold of within-group significance.

Working memory deficit is one severe symptom of vascular dementia patients such as aSAH survivors. Our study discovered significant decrease of characteristic path length, which indicates the increase of overall routing efficiency. We reason that patients with working memory deficit have to recruit more neuronal resources and thus develops higher overall routing efficiency of local network. This study provides novel information about the neural alterations of aSAH patients with working memory deficit. It might contribute to the understanding of its neural mechanism and the improvement of current intervention.

This study recruited limited number of subjects. However, the disease-triggered cognitive changes require larger sample size to determine.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by IRB of Huanggang Central Hospital. The patients/participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

CZ designed the study and wrote the article. R-SZ and TW analyzed the data. J-SZ reviewed this article. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.