Neurosurgeons commonly rely on locating the hand-motor cortex (HMC) to establish boundaries of functionally eloquent areas of the brain [1]. Establishing functional boundaries and sparing healthy tissue during presurgical planning is necessary to achieve the highest level of functional outcomes and potential recovery post-surgery. Presurgical mapping of eloquent cortex can be completed in many ways: using invasive techniques, such as direct cortical stimulation (DCS), and non-invasive techniques, such as task-based functional magnetic resonance imaging (tb-fMRI) and transcranial magnetic stimulation (TMS); however, there are limitations to these approaches.

The gold standard of hand motor localization has long been DCS, an intraoperative technique developed by Penfield in the 1940 s [2], [3], [4]. However, due to its highly invasive nature it is not a feasible option for presurgical planning [5], thus precludes insights into eloquent and functional cortex [2], [6], [7], [8], [9]. Moreover, DCS is both time consuming and risky: intraoperative fatigue, the potential to induce epileptic seizures, additional operative time and limitations created by the extent of the craniotomy are all factors that need to be considered when performing DCS [10], [11]. An alternative to DCS is the use of transcranial electric stimulation to elicit motor-evoked potentials (MEPs). It permits real time monitoring to subsequently minimize the risk of localized injury to neural correlates of the hand, and has good spatial resolution to allow for greater location information, and improve the outcome of treatments to this area. However, this stimulatory approach has some inherent limitations, such as false-negative results when stimulation occurs near but not in the motor cortex [12], [13], and variability in amplitude recordings resulting in the need for several time-consuming repeat trials, resulting in longer operative times [14]. These limitations support the utility of non-invasive techniques that provide presurgical insights and are time, and spatially-efficient.

Tb-fMRI is widely used in clinical settings with some studies demonstrating comparable sensitivity and specificity to DCS, however these results are varied and vascular changes have been proven to result in false negative fMRI results. This rendersthe method unreliable for tumor resection planning [15]; and has been criticized for its lack of reproducibility [16], [17]. Additionally, tb-fMRI requires a patient to be able to understand and conform to specific functional activities, a limitation for many clinical contexts. Despite the limitations, there are advantages to fMRI, particularly the non-invasive nature and repeatability, making it an valuable tool for presurgical mapping [11].

TMS, a more direct observation of functional cortex (relative to tb-fMRI), requires sophisticated and specialized expertize and equipment in order to produce clinically actionable insights [18]. There are many issues with the use of TMS for functional localization, mainly related to coil placement and limited precision of coil placement methods [19], [20] Craniometric methods of locating scalp targets and consequently brain regions are inherently inaccurate, thus image guidance-based methods have been developed to aid target accuracy [20], [21]. However, these advanced techniques are often not utilized in clinical practice due to the increased cost of buying the equipment [22].

Taken together, tb-fMRI, TMS, and DCS are (1) time consuming, (2) expensive, (3) and difficult to implement (i.e. require specialized equipment and expertize). Consequently, more sophisticated methods are necessary to decrease the significant limitations, resources, and risks associated with traditional methods, while facilitating post-surgical preservation of function.

The Structural Connectivity Atlas (SCA) [23] was developed as a brain parcellation method to overcome issues surrounding the mapping of abnormal brains in clinical contexts. The SCA relies on machine learning for precise and personalized brain mapping using diffusion weighted imaging (DWI). This enables preoperative planning and neuronavigation during surgery, regardless of individual anatomical variation [23]. The utility of the SCA within the Quicktome Brain Mapping platform (Sydney, Australia), which is an FDA-cleared tool for brain mapping,has been demonstrated in visualizing important functional networks and tracts, particularly in distorted brains [24]. However, its accuracy has not been investigated in identifying eloquent areas in a clinical setting.

The present study sought to validate the utility of the SCA in identifying the HMC, initially localized using TMS mapping. We used the visualization and modelling of the SCA paired with a high-resolution HMC from the Brainnetome atlas and examined whether the region identified matched with the HMC identified using TMS. The validation of this tool in a clinical setting may allow presurgical identification of eloquent areas to aid surgical planning and improve patient safety.