*3.5. Abnormal Dorsal Caudate Activation Mediates the Relationship between Information Processing Speed and Cognitive Flexibility in mTBI*

Mediation modeling was performed to test whether the magnitude of information about processing speed's effect on cognitive flexibility was dependent on reduced dCau activation. We found that reduced dCau activation had a significant mediating effect on the relationship between information processing speed and cognitive flexibility in mTBI patients (a ×b = −18.836, 95%CI: [−42.183, −5.347], *p* < 0.05, Figure 4).

**Figure 4.** Abnormal dCau activation mediates the association between information processing speed and cognitive flexibility in mTBI patients. The illustration demonstrates that within mTBI patients, information processing speed affected cognitive flexibility through the abnormal dCau activation. mTBI, mild traumatic brain injury; HCs, healthy controls; dCau, dorsal caudate; CI, confidence interval.

#### **4. Discussion**

To the best of our knowledge, this is the first study to investigate the underlying mechanisms of cognitive flexibility in mTBI patients using a rule-based task-switching experimental paradigm. The mTBI patients exhibited a longer RT than HCs on the TS condition, accompanied by a significantly reduced activation in the dorsal caudate, anterior cingulate cortex and other frontal regions. Furthermore, the abnormal dorsal caudate activation mediated the relationship between information processing speed and cognitive flexibility in patients. Together, these results suggested that dorsal caudate might play a vital role in the cognitive flexibility of mTBI, thus providing an alternative clinical target for impaired cognitive functions in patients.

During the task behavioral performance, mTBI patients only showed significantly longer RT on TS condition compared to HCs, indicating that mTBI patients had impaired cognitive flexibility on the behavior level. These data were in sync with data from previous studies [15,16]. These studies used local-global task-switching tasks and showed that, compared with HCs, patients with traumatic brain injury had significantly longer RT and increased ER under TS condition. Although the studies performed different experimental paradigms, the findings demonstrated specific cognitive behavioral patterns in patients with traumatic brain injuries. Previous research found that patients with several neurocognitive disorders exhibited poor cognitive outcomes [36,37]. We speculated that the mTBI patients had specific behavioral response patterns on impaired cognitive flexibility, which could help in understanding cognitive dysfunction in mTBI.

By comparing the brain activation response of mTBI patients and HCs under different task conditions, we showed that during the OG condition, mTBI patients exhibited significantly weaker activation in widespread frontal brain regions. Plenty of MRI studies revealed that the frontal cortex is involved in different neural circuits with a subcortical nucleus, which plays a vital role in various kinds of cognitive functions [38–41]. Many clinical studies have demonstrated that mTBI patients often present significantly reduced activation in the frontal cortex during simple tasks [42,43]. On the other hand, the observed abnormal frontal activation in mTBI indicated that the patients had cognitive dysfunction. In addition, during the DI condition, mTBI patients showed significantly reduced activation in the PCC, pIns and other frontal regions. As a key hub in the default mode network, PCC is involved in the regulation of attention and self-referential processing [44–46]. It has been hypothesized that reduced PCC activation in mTBI reduces the ability of other stimuli to direct patients' attention away from the distract stimulus, as patients showed significantly longer RT for the DI condition compared with the OG condition.

In the TS condition, our analyses showed that patients had significantly longer RT and higher ER compared to HCs, thus reflecting impaired cognitive flexibility in mTBI. Furthermore, compared with HCs, mTBI patients exhibited reduced activation in dCau, ACC and other frontal regions. Previous studies showed that the ACC, a hub region in the salience network [47], plays an important role in cognitive control [23], and has a close projection loop with the caudate nucleus [48,49]. Several studies performed the Stroop task, and showed that conflicts between potential behaviors elicit a significant abnormal activation during conflict monitoring and action selection [50–53]. In addition, the interconnection of the caudate nucleus and ACC is involved in evaluating the consequences of behavior and adjusting behavioral response [53–55]. Previous studies have shown that the caudate nucleus is a susceptible hub region in mTBI [56], and its dysfunction was significantly associated with impaired cognitive function in patients [57]. In addition, the impaired white matter fiber from the caudate nucleus to the cerebral cortex could predict impairment of cognitive flexibility in mTBI patients [16]. In our study, mTBI patients might have altered white matter connections in the dorsal caudate, which resulted in reduced activation in the patients. Another study showed that the abnormal caudate activation in patients with traumatic brain injuries was significantly associated with cognitive fatigue [14]. Our results suggested that mTBI patients might need more cognitive load to perform cognitive

flexibility responses, which would further cause cognitive fatigue, and finally lead to the abnormal activation of dorsal caudate [58].

Interestingly, we demonstrated that the abnormal activation of dorsal caudate played a significant mediating role between information processing speed and cognitive flexibility in mTBI patients. Many previous studies on traumatic brain injury have shown that structural and functional abnormalities of the caudate nucleus were significantly related to executive dysfunction in patients [14,57]. Besides, we showed that information processing speed was significantly related to cognitive flexibility in mTBI patients. This was consistent with the findings from previous studies, which showed that an individual's information processing speed ability is significantly associated with cognitive flexibility [59,60]. However, whether the impaired information processing speed or the abnormal dorsal caudate activation affects cognitive flexibility in mTBI patients remains unknown. Investigating the relationship between the information processing speed, abnormal dorsal caudate activation, and cognitive flexibility would help to understand the neural mechanisms of cognitive flexibility deficit in mTBI patients. Our mediation model showed that the abnormal dorsal caudate activation was a significant mediator between the patient's information processing speed and impaired cognitive flexibility.

There are some limitations that should be noticed. Firstly, it is still unclear whether the observed abnormal activation changes could be related to structural and resting-state functional alterations in mTBI patients. Future studies can combine multimodal MRI imaging to investigate the neural mechanism underlying cognitive flexibility in mTBI. Secondly, the difficulty level of the experimental task in this study was relatively high for patients. Thus, future studies can apply cognitive tasks with a moderately difficult level to investigate cognitive function in mTBI. Finally, the sample size of both groups was relatively small. Future studies with larger samples are needed to improve the reliability of the results and investigate the effect of different brain injury types on results.
