**4. Discussion**

In this study, we attempted to provide empirical evidence on the question of whether DBS may induce psychological changes using an empirical, inferential approach with patients diagnosed with idiopathic PD. To this end, we targeted two specific variables, namely illness representations and coping strategies.

Based on the existing literature, we had expectations of various kinds of change that DBS would bring in as formulated in two research hypotheses. Yet, globally, our results did not provide confirmation of these expected changes. The first hypothesis, related to illness representations, was partly invalidated as PD was perceived as less cyclical after surgery than before, although patients had lower representations of personal control on PD post-surgically, this latter finding being expected. Regarding the second hypothesis, we observed that the use of instrumental coping strategies was not modified by DBS, which invalidates our initial assumption.

These results sugges<sup>t</sup> that PD remains globally perceived as the same medical entity after DBS, despite the important reduction of motor symptoms induced by brain surgery. Patients' representations of control on PD decreased after surgery, and this was notably the case for personal control. The notion of control is ethically important as it is associated with autonomy, which, in this context, relates to the patient's right to decide about his/her treatment [26]. In this regard, feelings of self-estrangement reported from testimonies of operated patients were associated with the impression of losing the control over one's emotions and capabilities [48]. More generally, a loss of personal control in illness managemen<sup>t</sup> was expected as it has been well illustrated by testimonies of operated patients [12,49]. In terms of adjustment, this is problematic as subjective perception of control over a chronic disease was associated with adaptive outcomes [33]. This association was notably established in PD patients not treated with DBS [50]. It was however surprising to observe a diminution in representations of treatment control of PD—although this was weak and suggests that, globally, perceptions of treatment control remain similar before and after surgery. One might have indeed imagined that the necessity to rely on neurologists to adapt stimulation parameters would have led to an impression that health professionals may manage the disease e fficiently. Possibly related to this latter issue, patients had a less cyclical representation of PD after DBS than before. This suggests that regular appointments with a neurologist to adapt stimulations parameters are not su fficient to create a cyclical representation of PD; however, stabilization of motor fluctuations induced by DBS surgery [51] may have instilled the feelings of a medical condition having become less cyclical after surgery. Representations of PD as becoming more stable after DBS may be related to better psychological well-being, as it was measured in non-DBS patients diagnosed with PD [52].

Finally, we found that use of instrumental coping strategies to deal with stressful situations was not reduced after DBS. This result, which adds to some previous findings [41] but stands in contradiction to others [14,39,40], suggests that DBS does not deeply modify strategies of stress management, the latter being not exclusively dependent on situational issues (e.g., intensity of motor symptoms before vs. after DBS).

Overall, the findings of this study highlight the stability in illness representations and coping strategies throughout the DBS process and rehabilitation. Although significant changes were found here and there, their magnitude was generally small. The only notable exception was representations of PD as being less cyclical after DBS than before. Yet, even in this specific case, magnitude of change should be interpreted with caution. For instance, the implied mean at T0 retrieved from the free model was 13.515 and that at T1 was 11.779. As representations of cyclical timeline are assessed in the IPQ-R with four items, these means correspond to an item mean score of 3.379 at T0 and 2.945 at T1. In other words, representations of cyclical timeline remain around a mean item score of 3 corresponding to a "neither agree nor disagree" response anchor at each measurement time. Thus, even significant and with a medium to large e ffect size, changes in representations of cyclical timeline remain globally stable over the DBS process and should not be associated with a major alteration in patients' perception of PD. Interestingly, these small changes were all identified by a critical examination of both FIML and

Bayesian estimations in SEM, a method that does not suppress missing data from analyses; on the contrary, rANOVA analyses did not find any significant difference between measurement sessions. Methodologically, this observation implies that SEM remains discriminant with small-size samples, which is a scenario likely to happen frequently in heavy medical situations such as DBS surgery.

Our results should be considered with caution because of limitations inherent to the study design. First, illness representations and coping strategies are not representative of all psychological aspects of the DBS experience. Besides, we decided to focus our analyses on a limited number of illness representations and coping strategies, based on assumptions from the literature and methodological limitations of our experimental design; it remains nevertheless possible that other kinds of illness representations and coping strategies are significantly altered after DBS. Finally, the findings of the current study stand for patients undergoing DBS for PD; it cannot be ruled out that individuals treated with DBS for other medical conditions would experience different outcomes regarding the variables investigated in this research.
