*4.3. Limitations*

Our study had some limitations. First, it was a retrospective study, involving patients for whom there was no standardized PSA protocol and uncontrolled variability in PSA drug choices and doses. We found that demographic and clinical characteristics did not di ffer systematically between the groups, therefore these factors could be accounted for with a regression analysis thanks to our large overall sample size. Since no standardized PSA protocol was used we were limited when testing for dose-dependent e ffects. Second, while we compared our findings to previous reports, the available

literature is sparse and characterized by heterogeneity in PSA protocols, patient characteristics and MER data analysis. Our findings sugges<sup>t</sup> that some di fferences among previous reports may be accounted for by dose-dependent e ffects (e.g. on firing rates). Prospective studies with standardized PSA protocols are required to confirm these findings. Also, we did not assess spectral changes in the local field potential. Investigating the e ffect of PSA on the presence of pathologic oscillations remains unanswered. Another limitation is that this study did not assess the e ffects of the various agents on intraoperative clinical measures such as tremor. A follow-up study is needed to assess the e ffects of the various PSA protocols on STN depth and size. Moreover, it is important for clinical practice to address in future studies whether the use of PSA influences clinical outcome of STN DBS.
