In this work, we address the need of accessible and comprehensive analytical tools for the DBS therapies. We present a toolbox with an intuitive user interface that allows both clinicians and researchers to access and explore intricate electrophysiological data. To the best of our knowledge, no other toolbox offers a comparable range of functions in terms of both extent and variety, together with ease to run/setup and customize (but see ref. 17). Among these functions, we would like to emphasize the inclusion of an ECG cleaning tool and the capability to visualize streaming recordings with data obtained from wearable devices. Furthermore, the toolbox operates without the need for an internet connection and patient data is not required to be stored on a server. While we acknowledge that a central database may be desirable in a healthcare setting, it is also true that it imposes additional technical/security measures regarding the sharing of personal data.To standardize some of the most performed analyses of the field, we designed a novel structure of classes and developed a diverse repertoire of functions within each class. Some of these functions were adapted from existing toolboxes, notably in the detection and cleaning of cardiac artifacts. By adopting an object-oriented programming framework, we were able to produce a toolbox that can analyze multiple files in a simple and intuitive workflow. This is desirable not only from a software design perspective, but also from a scientific perspective, where longitudinal, multi-signal and multi-patient analyses are relevant11,19,20.To illustrate the application of DBScope, we present two case studies that concentrate on out-of-clinic and in-clinic recordings. The latter are more commonly used in the context of Percept PC, as they offer a higher temporal resolution. Nevertheless, out-of-clinic recordings still provide an invaluable glimpse into real-world settings, where data-driven DBS therapies aim to actuate. In fact, we observed that chronic data contains valuable information, such as the patients’ circadian patterns, which exhibit longer temporal dynamics. These patterns are challenging to identify in in-clinic recordings but are discernible in long-term ones. At the same time, it was possible to study the response of the LFP to specific occurrences, such as medication intake or rigidity episodes. While the marking of medication episodes showed potential in monitoring the wear-off effect, other events were reported less times and typically coincided with medication intake, suggesting that the patient was reporting retroactively. Regarding the in-clinic recordings, the toolbox offered clear visualizations that facilitated the identification of both artifacts and clinically relevant frequency bands. Additionally, the possibility of aligning streaming recordings with wearable data proved to be a relevant addition to DBScope, enabling the investigation of movement-related modulations.Although DBScope succeeded on many fronts, it bears some limitations. Presently, the toolbox operates exclusively on files extracted from Medtronic’s Percept PC setup. As new devices with sensing capabilities have since emerged (Medtronic’s Summit RC + S and Newronika’s AlphaDBS), we are considering the development of specific parsers to accommodate them. DBScope is also reliant on commercial software (MATLAB). However, it should be noted that universities and research institutes often offer campus-wide licenses to their members.Future developments aim to address two challenges. The first lies in the search for biomarkers. Defining a biomarker, or library of biomarkers, is an essential, albeit complex, step for the development of patient- and symptom-specific DBS therapies21,22,23. A direct mapping between brain activity and reported symptoms is seldom possible, due to the subjective and elusive natures of the latter. In these cases, a second signal, more interpretable and highly linked with the symptom type, is often introduced to facilitate the mapping process. For instance, accelerometry data is widely used to link electrophysiological signals with motor symptoms. In light of this, we have already included the functionality to load accelerometry data into DBScope and are actively developing additional methods to leverage the information contained in these signals. The second challenge stems from the fact that LFP are prone to artifacts of different origins, such as cardiac, movement, and stimulation. Although current devices have implemented artifact-cleaning algorithms, not all artifacts can be reliably identified and effectively cleaned. DBScope currently enables artifact screening in an iterative process, where the user alternates between visualization and filtering/cleaning steps. However, this approach is time-consuming and ineffective when the entire frequency spectrum is affected. In this respect, we are invested in the development of algorithms that not only automatically detect the source of these artifacts, but also clean them accordingly.One of our major goals was to create the conditions for the DBS community to adopt and easily contribute to the enhancement of this toolbox, in response to advancements and discoveries in DBS research. For this reason, DBScope is “open source” and is accessible through an online repository. The dependence on MATLAB was circumvented with the creation of standalone applications for both Microsoft Windows and macOS operative systems. These standalone applications, also available in the online repository (“Release” section), allow DBScope to be used without having to install MATLAB (royalty-free). We firmly believe that these types of initiatives are key in fostering the emergence of novel methods for clinical integration. Although we are currently working on updates, we encourage the clinical and research communities to adapt the tools and algorithms already available in DBScope and to share their insights, becoming part of this joint effort to improve the DBS therapy.In conclusion, DBScope is an open-source computational toolbox to import, visualize and analyze files from the Percept PC device (Medtronic, BrainSense Technology). The toolbox can be used programmatically or through an interface for users without programming experience. This way, it can be directly integrated into the clinical and research practices, whilst remaining adaptable to new research questions. Its functionalities are up to date with current literature standards for the evaluation of LFP in DBS. Moreover, future updates are in store. Overall, DBScope is a versatile tool focused on the widespread improvement of data-driven DBS therapies, by expanding the accessibility of the data and by promoting new forms of analyzing the complexity of DBS data. DBScope is available for download in GitHub, and can be accessed via this link: https://github.com/NCN-Lab/DBScope.
