Ires Ghielen

172 Chapter 9 patients on based on the total score on the BAI (cut-off >12) [42]. This might have resulted in some false positive associations between BAI-items in the network. Although the patient groups were relatively large (316 and 253 PD patients) and we corrected for multiple comparisons, for this type of analysis it is preferred to test even larger groups [43]. It is therefore important to interpret the results of this chapter with caution. The described BEWARE training in chapters 7 and 8 and the studies included in the meta-analyses described in chapter 6 included small numbers of participants. This results in low power, and therefore we cannot draw hard conclusions concerning the effectiveness of psychological treatments. Although the psychological treatments and the BEWARE training show promising results, there is room for improvement. Suggestions for future research This thesis elaborately investigated the complex interactions between motor and non-motor symptoms in PD patients, including the treatment of anxiety in the context of response fluctuations. However, more research is needed to fully understand these complex interactions to improve diagnosis of anxiety and wearing-off related distress and its treatment. Since these complex interactions are very much apparent in PD patients with fluctuations, we expect that in these patients the interactions might be even more evident when symptoms are measured in shorter time periods (e.g. minutes or hours). For example, a panic attack occurs suddenly [8] and usually diminishes within several minutes. PD patients that experience panic attacks in the context of wearing-off related anxiety can experience subsequent symptoms that occur in a time-period of minutes. To further investigate symptom interactions, it is therefore recommended to apply methods with more finegrained time resolution to study interactions between motor, automic and affective symptoms of response fluctuations. An interesting example of such approach is the N = 1 study performed by van der Velden and colleagues [44], investigating time-series data from an individual patient using network analysis. Building onto our own research, it would be interesting to include both somatic symptoms (such as freezing, tremor, rigidity, and autonomic symptoms) as well as cognitive of affective anxiety symptoms (such as fear of losing control and rumination). We hypothesize that, in anticipation of an ‘off’ period, patients feel more tense and therefore experience an earlier onset of or a more severe subsequent ‘off’ period. In addition, by using time-series data in a network analysis for a single PD patient, we can investigate which symptoms interact with other symptoms and are therefore potential treatment targets. This might