Study of rapid reorganization of visual neurofunctions with the resting‐state functional MRI in pituitary adenoma patients with vision improvement after transsphenoidal surgery

Abstract Introduction To investigate changes of vision‐related resting‐state activity in pituitary adenoma (PA) patients with visual improvement after transsphenoidal surgery. Methods 14 PA patients with visual improvement after surgery were enrolled. The resting‐state functional MRI and neuro‐ophthalmologic evaluation were performed before and after the operation. The functional connectivity (FC) of 8 seeds (the primary visual cortex (V1), the secondary visual cortex (V2), the middle temporal visual cortex (MT+), and fusiform gyrus(FG)) was evaluated. A paired t test was conducted to identify the differences between the two groups. Results Compared with the preoperation counterparts, the PA patients with improved vision exhibited decreased FC with the V1, V2, MT+, FG in the left paracentral lobule, bilateral lingual gyrus, precentral gyrus(BA 4), right superior temporal gyrus(BA 22), left fusiform gyrus, bilateral middle occipital gyrus (BA 19), left cuneus, right inferior occipital gyrus, left superior frontal gyrus, right cuneus, left superior parietal lobule(BA 7),the medulla, right postcentral gyrus, and increased FC in the right middle frontal gyrus, left inferior parietal lobule (BA 40), left declive, right lentiform nucleus, inferior frontal gyrus, right superior frontal gyrus(BA 11), cingulate gyrus(BA 32), right putamen, right thalamus, left medial frontal gyrus, left claustrum, left superior frontal Medial, right rectal gyrus(BA 25) and right parahippocampal gyrus. Conclusions The results show most subareas within the visual cortex exhibit decreased functional connectivity. The functional changes in subareas within default mode network (DMN), action observation network (AON) and the multisensory system in PAs propose that vision improvement may lead to function remodeling in higher‐order cortex.


| INTRODUC TI ON
Neuroplasticity means the inherently dynamic biological ability of the central nervous system to develop maturation, to reshape structurally and functionally in related to experience, and to adapt after injury. The human visual cortex is the good tool to study the development and plasticity of the neocortex. The experience-dependent neural plasticity is mainly present in the developing visual cortex.
The human visual cortex demonstrates immature at birth, develops the deep perception at about 6 months after birth, and acquires maturation in late childhood (Kovács, 2000). However, the potential plasticity is presented in the adult visual cortex .
Therapeutic interventions can also trigger plastic changes in the aging visual cortex by restoring vision (Lou et al., 2013).
The visual cortex comprises a primary visual field (V1) and numerous extrastriate/association visual areas (Felleman & Van Essen, 1991). The innervation of the V1 by the thalamus is the lateral geniculate nucleus (LGN), and the middle temporal cortex (MT+) is the inferior pulvinar (PIm) and LGN . These retinothalamic pathways to area MT + could be important drivers/modulators of visual perception in the adult (Laycock et al., 2007).
The two-stream hypothesis is an influential and widely accepted model of visual information processing (Ungerleider & Mishkin, 1982). From V1, the dorsal stream (the "how pathway") extends to area MT+ (V5), and on to the posterior parietal cortex, in which way spatial/motion vision was processed. The ventral pathway (the "what pathway") continued to V4 and terminated in the inferior temporal cortex, in which way form/object vision was processed. The ventral stream pathway included areas with strong preference for faces in the fusiform face area and occipital face area, for body parts in the extrastriate body area. Many of the areas between two parallel streams are very likely interconnected (Braddick et al., 2000;Farivar et al., 2009). "Top-down" refers to cognitive influences and higher-order representations that impinge upon earlier steps in information processing. The top-down signal can facilitate the integration of objects and backgrounds in the visual scene to make a stable representation of the objects within it (Gilbert & Li, 2007).
FMRI is an indirect measure of neuronal function by measuring localized changes in the oxygenation of blood hemoglobin. It is noninvasive and well tolerated by patients. Resting-state functional magnetic resonance imaging (RS-fMRI) is used to evaluate functional connectivity (FC) between brain regions when patients are at rest (Fox & Raichle, 2007;Pillemer et al., 2017). FC refers to the temporal correlation between spatially remote neurophysiological events (Friston et al., 1993). FC analyses based on BOLD signal are particularly promising because they can offer high spatial resolution and high spatial specificity relative to where the corresponding changes in neurophysiological signals take place compared with all other noninvasive imaging modalities (Shmuel et al., 2007). FC analyses can potentially further understanding of neuroanatomical models (Fox & Raichle, 2007).
Pituitary adenoma (PA) may compress the optic chiasm, optic nerve, or optic tract. The patients with PA often present with impaired vision. The transsphenoidal surgery is minimally invasive and does no harm to the visual cortex and other closed region. After the removal of tumor, many patients' vision was improved. In those PA patients who had improved vision at about 3 months after the operation, the results showed that regional homogeneity (ReHo) decreased or increased within the visual cortex and some brain region (Qian et al., 2015). In clinic, some patients can acquire improved vision right away after surgery. During the early time after surgery, however, the change of neurofunction in visual cortex and higher cognitive cortex is still yet to be explored so far. In literature, only one study presented the neurofunction change of vision areas after visual restoration after 3 days following surgery (Giulia et al., 2012).
But in this research, only one patient was recruited, and the task-fMRI was used to evaluate the brain function. So we recruited more PA patients with improved vision at about 3 days after transsphenoidal operation. We used a priori defined regions of interest (ROIs) in areas V1, V2, MT+, and FG (fusiform gyri) to analyze RS-fMRI data. Our study is to evaluate the changes in the vision-related resting-state network in PA patients. Furthermore, we plan to explore the plasticity of some specific subareas within the visual cortex and higher cognitive networks after restoration of vision.

| Subjects
16 patients with pituitary adenoma were enrolled in this study. All those patients with the visual damage underwent transsphenoidal tumor resection surgery and got restored vision right away after the surgery. PA patients were recruited according to the inclusion criteria: Age ranged from 18-65 years; the corrected vision acuity was below 1.0 (20/20) before the operation, ophthalmologic diseases or other intracranial lesions that affected the visual pathway or cortex were ruled out; vision improvement at the 3 day after operation (the corrected vision acuity improved by more than 0.2 at least unilaterally) was required; and no severe electrolyte disturbance, hypopituitarism or other complications presented after the operation. This study was approved by the Ethics Committee of Hospital. Written informed consent was obtained from the patients.

| Data acquisition
Images were acquired one day preoperatively and three days postoperatively on a 1.5 T MR system (Espree, Siemens Medical Solution, Erlangen, Germany) in the diagnostic room of iMRI brain suite, which was described in detail previously (Chen et al., 2012). Foam pad was used to minimize head movement, and earplugs were set to decrease acoustic noise during scanning. During the RS-fMRI scan, the patients were instructed to remain motionless and keep their eyes closed and not to think systematically. RS-fMRI data were acquired using an echoplanar image pulse sequence (slice thickness = 4.5 mm, flip angle = 90°, and field of view [FOV] =224 mm × 224 mm, TR: 2,000 ms TE: 45 ms.)

| Clinical and neuro-ophthalmologic assessments
We evaluated the cognition of all patients using the mini-mental state examination prior to the operation. The patients underwent neuro-ophthalmologic examination within 2 days prior to the operation and at approximately 3 days after the operation. We measured the best-corrected visual acuity for distance with the E chart and made report in the decimal scale. We performed the ophthalmic fundus examination with a nonmydriatic retinal camera (Topcon, Japan).

| FC and statistical analysis
First, noise-related variance, including six head motion parameters, the global mean signal, the white matter signal, and the cerebral spinal fluid (CSF) signal, was removed from the preprocessed data by linear regression analysis. The images were then spatially smoothed with a 6-mm FWHM Gaussian kernel. The individual FC maps were transformed to z-maps to improve data normality. A paired t test was performed on the z-maps to show significant differences in correlation between the two groups. The AlphaSim method was selected to correct for multiple comparisons. The corrected value of p < .05 (uncorrected p < .001 and a minimum of 40 voxels in a cluster) was used as the threshold.

| Studied population
According to the inclusion criterion, 16 patients were recruited in our study. As a result of head motion or the lack of sufficient data after scrubbing, 2 patients were excluded; thus, 14 patients (male/ female 7:7) were included in the final analyses. Mean age was 46.3 (range 24-62 years). The main demographic and clinical characteristics of patients are listed in Table 2.

| Ophthalmologic evaluation
Detailed results of the ophthalmologic evaluation are reported in Table 2.

| Decreased FC in the patients after operation
Compared with the preoperative counterparts, decreased FC with left V1 was identified in the left paracentral lobule ( Figure 1, Table 3).
Decreased FC with right V1 was identified in the right lingual gyrus and left precentral gyrus (brodmann area (BA) 4) ( Figure 2, Table 3).
Decreased FC with left V2 was identified in the right superior tem-   Figure 3, Table 4).
Increased FC with right V2 was identified in the right superior frontal gyrus (BA 11) and right cingulate gyrus (BA 32) ( Figure 4, Table 4).
Increased FC with left MT + was identified in the left cingulate gyrus ( Figure 5, Table 4). Increased FC with right MT + was identified in the right putamen, right inferior frontal gyrus, right middle frontal gyrus, right thalamus, left medial frontal gyrus, left claustrum, and left superior frontal medial ( Figure 6, Table 4). Increased FC with left FG was identified in the right rectal gyrus (BA 25) and right parahippocampal gyrus (Figure 7, Table 4). after visual restoration in PAs.

| D ISCUSS I ON
The dorsal stream originates from the V1 to V2 to MT+, and arrives at the inferior parietal lobule. This pathway is related with the detection of motion and location, and the control of the eyes and arms (Merigan & Maunsell, 1993;Tootell et al., 1998  More fMRI studies are needed in the future. The main visual information is relayed from the retina to the LGN to V1 (BA17) to V2 (BA18), and into higher-order visual cortex. Some minimal visual information is transmitted from the retina to the pulvinar and LGN and directly to MT+/V5 (Lyon et al., 2010;Warner et al., 2010), an alternative pathway that bypasses V1. Previous studies showed that MT + may be a potential substitute when a lesion in V1 occurs at early age (Bridge et al., 2008;Werth, 2006). The thalamus-MT bypass may play a compensatory role in the vision loss because of anterior vision pathway diseases (Mascioli et al., 2012).
The retinothalamic pathways to MT + could play an important role in the drivers/modulators of visual perception in the adult (Laycock et al., 2007). Our data show that FC between thalamus and MT in- Compared with the preoperative counterparts, increased FC with V1, V2, and MT was identified in the right putamen, right lentiform nucleus, right thalamus, and left claustrum. The right putamen, right lentiform nucleus, right thalamus, and left claustrum are subareas of the multisensory system (Cappe et al., 2009). The multisensory system at the cortical locations includes the parietal lobe, temporal lobe, frontal lobe, and insular. The multisensory system at the subcortical locations includes the superior colliculus and basal ganglia (globus pallidus, caudate nucleus, putamen nucleus, amygdaloid body, claustrum nucleus) (Brown et al., 1997;Cappe et al., 2009).
The sensory-specific thalamic structures play an important role in multisensory integration processes and behavior performances (Tyll et al., 2011). The putamen can integrate the neuronal interactions between visual recognition and articulatory areas (Seghier & Price, 2010). The claustrum nucleus connects with the visual cortex and integrates the information between the earlier visual cortex and vision-related thalamic nucleus (Olson & Graybiel, 1980). The  (Caspers et al., 2010). Observing others' actions causes reaction in many sensorimotor cortexes that collectively consist in a network named the AON (Cross et al., 2009;Gazzola et al., 2007). The AON is proposed to contribute to the understanding of others' actions by coping those actions into one's own motor system (Rizzolatti & Craighero, 2004 , 2006;Ortigue et al., 2010), which has been linked to social cognition (Cross et al., 2009;Kaplan & Iacoboni, 2006;Sobhani et al., 2012). The AON's functional and structural connectivity at rest has been related to behavioral measures of social and motor skills (Fishman et al., 2015;Williams et al., 2017). Our results showed the decreased connection between the visual cortex and substrate of AON after vision restoration. The AON may be involved in the neural reconstruction for vision recovery.
In the FC analysis, we identify increased FC with ROIs in the bilateral cingulate gyrus (BA 32), and left medial frontal and superior frontal medial gyrus after visual restoration in PAs. All of these areas are subareas of the default mode network (DMN). The DMN plays a role in the detection and monitoring of both environmental events and internal mentation (Rudebeck et al., 2013) and mediates subject responsiveness and the saliency of external stimuli (Andrews-Hanna et al., 2010;Leech & Sharp, 2014;Rudebeck et al., 2013;Wen et al., 2013). When the DMN detects the decreased visual cortical activity, the decreased deactivation in DMN may likely occur. The strong DMN activity is related with reduced visual cortical excitability (Mounder et al., 2013).
Our results show that the FC decreased in visual cortex after vision recovery, so it may be justified to propose that decreased visual cortex activity in some way incur decreased DMN deactivation (stronger activity). However, the mechanism resulting in functional alteration in DMN after vision restoration is still to be elucidated.
Increased FC with FG in the right rectal gyrus (BA 25) and right parahippocampal gyrus mean increased neural connection after vision restoration. The rectal gyrus is subareas of the ventromedial prefrontal cortex (vmPFC). The vmPFC has been implicated in a variety of social, cognitive, and affective functions. The vmPFC modulates facial emotion recognition through its interactions with posterior cingulate cortex, precuneus, dorsomedial prefrontal cortex, and amygdala (Hiser & Koenigs, 2018). Our results propose that the increased response of rectal gyrus is related with the neurofunctional reorganization after the vision restoration. The parahippocampal gyrus is located in the inferior temporooccipital cortex, surrounding the hippocampus. The parahippocampal gyrus is involved in visual scenes (Mégevand et al., 2014), cognition (Aminoff et al., 2013), and spatial control (Aminoff et al., 2007). The parahippocampal gyrus has been suggested to control the processing of object and scene information (Staresina et al., 2011). A previous study showed the activation of the parahippocampal gyrus when a three-dimensional spatial structure was presented (Henderson et al., 2008). Our results suggested that the vision restoration might enhance the function of the parahippocampal gyrus.
We also found an increase in the FC between the V2 and the cerebellum (declive). The cerebellum, which functionally interacts with the frontal eye fields (Kelly & Strick, 2003;Middleton & Strick, 2001), is also involved in the control of eye movements (Hayakawa et al., 2002;Nitta et al., 2008). Damage to the cerebellum can affect smooth pursuit eye movement (Straube et al., 1997). Our data proposed that the vision improvement leads to the increased function of the cerebellum. we will make well-designed experiment to explore changes in brain function in PA patients with vision restoration.

| CON CLUS IONS
Collectively, we showed rapid reorganization of neurofunctions in the vision-related cortex of PA patients with visual improvement.
Most subareas within the visual cortex exhibited decreased FC.
The MT + exhibited enhanced FC with the thalamus, which may indicate an important role in the compensatory mechanism following visual improvement. The functional changes in subareas within DMN, AON and the multisensory system in PAs proposed that vision improvement may lead to function remodeling in higher-order cortex beyond the visual cortex. However, more studies are needed to explore the mechanism of neural plasticity within the visual cortex, as well as the mechanism of the interaction between the visual and higher-order cortex in patients with specific visual diseases.

ACK N OWLED G M ENTS
The authors wish to acknowledge Dr Yanyang Zhang for his help in interpreting the results of this study, and Dr Jiashu Zhang for his help in writing the paper.

CO N FLI C T S O F I NTE R E S T
The authors have no conflicts of interest to declare.

PEER R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1002/brb3.1917.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.