Concerning the particular case of cognitive rehabilitation in DLGG patients, we have to keep in mind the slow-growing and infiltrating character of DLGG, which make their associated cognitive disorders particularly amenable to rehabilitation. Indeed, on the one hand, by infiltrating cortical and sub-cortical structures, the tumor may destroy some of them but also only displace others and then residual function may be maintained [3, 38]. On the other hand, by growing slowly, the tumor induces a reactive reshaping and reorganization of functional networks [39]. It is precisely this key feature that may explain why cognitive or neurologic disorders can be more easily compensated as one goes along the disease compared to acute events like stroke [6]. In this respect, neurophysiological studies are really informative. Studies using functional magnetic resonance imagery (fMRI) paradigms have demonstrated different patterns of functional reorganization at the cortical level, showing that the brain recruits alternative areas not previously implied for the expression of cognitive function. Among these functional strategies, ipsilateral, perilesional, as well as homonymous contralateral recruitments have been described (for a review, see [6]). In this context, DLGG offers a unique and exciting opportunity to better understand both the dynamics underlying functional plasticity and the neural implementation of cognitive processes, valuable data for cognitive rehabilitation.

Thus, cognitive rehabilitation might on the one hand enhance residual functional capacities, and, on the other hand, potentiate the spontaneous functional brain reorganization.

The dynamic and holistic organization that assumes functional plasticity finds its corollary in the studies of cerebral connectivity in normal brains. For over than 10 years, new technics of data analyses more and more sophisticated, from functional and morphologic imaging, have emerged. In this setting, the idea according to which the brain is composed of complex large-scale neural networks became dominant. However, this view was already present in the middle of last century with Daniel Hebb [40] who suggested that high-level human functions are determined by the activity of complex neural networks composed of local and distant areas across the whole brain.

Data from spatial reconstruction of anatomical connectivity by means of diffusion tensor imaging, a technique that measures the diffusion of water molecules through cerebral tissues, is perhaps the best illustration of this complexity. The visualization of connections between distant brain areas via the multiple white matter bundles (projection or association fascicles, U-shaped fibers) is really demonstrative about this organization [41]. Although these data are by nature anatomic and give no direct functional information, studies using direct electrical stimulations during awake neurosurgery [39], which induce transient disconnection syndrome, have proved their essential role for the complete and normal expression of functions. In neuropsychology of strokes, injury of these subcortical fascicles can provoke severe cognitive disturbance, hardly compensable [42, 43]. In this context, it has been proposed that these subcortical structures are crucial for functional plasticity [44, 45]. If so, we should find structural changes of these white matter fascicles in reaction to tumors and neurosurgery, and correlate their markers (i.e. fractional anisotropy and mean diffusivity) to cognitive functions. However, to the best of your knowledge, there is for the moment no study with DLGG patients in which structural integrity and changes have been assessed in a systematic manner by the means of a longitudinal design (pre-/postsurgical). As a consequence, this type of structural plasticity remains to be demonstrated in the framework of this brain pathology. Yet, recently works in the field of the surgery of refractory temporal epilepsy are very interesting in this respect. For example, the team of Duncan [46] has found, using a pre-/postsurgical design, DTI, and language assessment, a correlation between the score obtained in verbal fluency after the surgery, and an increase of fractional anisotropy in several regions including subcortical structures such as corona radiata. In other words, these results show that the degree of language recovery is related to structural changes implying some white matter pathways. These provocative data, by demonstrating for the first time that we can call functional subcortical plasticity, open exciting perspectives in the field of brain tumors.

In addition to the fact that DTI can be combined with functional data like cognitive scores, other methods are particularly promising, especially those implying functional connectivity computing, to track the phenomena of plasticity induced by the tumor and its resection. Functional connectivity (FC) is defined as “the correlation between spatially remote neurophysiological events” [47]. This means that temporal statistical interdependencies can be found between several cortical areas composing the neural networks sustaining cognitive functions.

In several studies, abnormality of FC has been correlated with cognitive disorders, demonstrating that temporal desynchronization (hypo- or hyper-synchronization) between distant brain areas is particularly deleterious for functions. It is for example the case in neurodegenerative diseases where several and distinct patterns of functional alteration within the different networks can be found and linked to neuropsychological phenotypes (for a review, see [48]). For example, memory loss has been related to FC decrease in Alzheimer’s disease [49]. In the field of brain tumor and traumatic brain injury, several works have studied the impact of brain damage on FC (see Chap. 21). Bartolomei and colleagues [2] have shown using resting MEG (magnetoencephalography) paradigm that synchronization was altered in a population of patients harboring a brain tumor. In a subsequent study, using the same experimental design, cognitive disturbances were shown to be correlated to abnormality of FC [50]. Very recently, FC-based resting MEG at the level of the tumor was evaluated before brain surgery. It was found that decrease resting-state FC was highly predictive to the lack of functionality of this region as evaluated by means of direct electrical stimulation during awake surgery, suggesting that FC is a good measure of the integrity of brain functions [51].

In the same vein, patients with TBI show altered FC [52]. Nakamura and colleagues [53] demonstrated using a resting-state fMRI that just after the injury rsFC was disturbed and that, during recovery, this disturbance tended to normalize. A more recent study showed that cognitive complaints were predictive of altered FC in the default mode network in semi-acute TBI patients [54]. Furthermore, Castellanos and his team [55] have evaluated FC-based rsMEG in a population of TBI patients. Data were recorded immediately following the traumatic event and after a specific cognitive rehabilitation program. The authors found that neuropsychological performances significantly improved after treatment. Interestingly, this cognitive recovery was correlated with the reorganization of neural networks as indexed by the comparison between the pre- and posttreatment. These results suggest that functional recovery is related to the reorganization/reconfiguration of neural networks.

Of course, the main goal we aim to reach when proposing a cognitive rehabilitation is to bring the patient to recover a satisfactory level of cognitive functioning. The question is as follows: what is a satisfactory level of cognitive functioning? We think that there is no unique answer to this question and that it depends on the patient, his personality, and his expectations. Thus, the program of cognitive rehabilitation has to be established taking into account not only the objective assessments of cognitive functioning but also the subjective complaints and the expectations of the patient. In this state of mind, we approve and recommend applying the following proposal, borrowed from Kurt Goldstein’s works [56–58], a precursor of great influence in cognitive rehabilitation:

The relevance of taking into account the patient in his wholeness, and not only in his cognitive functioning, has been confirmed in a more recent study [59].

Moreover, the patient has to be informed about our objectives and how we project to reach them, in order to establish a real therapeutic alliance.

Taken together, the observations mentioned above suggest that brain injuries impact the functional coupling and integration between distant brain areas and that this alteration can be related to cognitive disorders. However, some results in studies with TBI patients show also that the spontaneous reorganization of neural networks is correlated in some extent with cognitive improvement. More interesting, cognitive rehabilitation with significant functional outcomes helps the brain to reorganize its functional networks [54]. Although these seminal results remain to replicate, they are particularly promising for cognitive rehabilitation. Understanding the dynamic of neural network reorganization and the optimal functional reconfigurations may have several important implications for the elaboration of cognitive rehabilitation strategies (e.g. which networks should be targeted).