The disruption of normal patterns of structural brain connectivity is believed to play a central role in the pathophysiology of many neurological and psychiatric disorders, such as dementia, movement disorders, stroke, traumatic brain injury (TBI), etc. White matter (WM) pathways consist of myelinated axonal structures that constitute the connectivity between different brain regions. Axonal injury and degeneration may occur even in the absence of tissue disruption. Therefore, patients may experience significant impairment despite the absence of abnormal findings on conventional CT or MRI. Moreover, focal imaging abnormalities that can be detected using CT and MRI are poor predictors of outcome (Niogi and Mukherjee, 2010; Adalbert and Coleman, 2013; Wang et al., 2016; Salvadores et al., 2017).

Axonal injury is a key determinant of clinical outcome in cases of brain injury and has been shown to be an important factor determining long-term motor, cognitive, and neuropsychiatric disability following brain injury (Mac Donald et al., 2007; Di Paola et al., 2010; Moura et al., 2019). Diagnostic tests that can discriminate significant axonal injury and degeneration are needed in order to accurately assess injury severity and effectively determine treatment and follow-up pathway.

WM pathways determined through diffusion tensor imaging (DTI) are traditionally considered to be the biophysical representation of axonal bundles and their myelin sheets. DTI streamlines between cortical and subcortical gray matter (GM) regions of interest (ROIs) can be used as a measure of the magnitude and strength of connection between ROIs (Basser et al., 1994; Le Bihan et al., 2001). Analyzing changes in brain connectivity using DTI tractography is widely used to evaluate axonal injury that is a hallmark of stroke and TBI (Arfanakis et al., 2002; Lo et al., 2010), in which WM tracks can be injured directly or indirectly through Wallerian degeneration (the anterograde distal degeneration of injured axons accompanied by demyelination) (Chen et al., 2017; Moura et al., 2019). DTI metrics are used to address specific or diffused WM damage that is frequently affected by stroke lesions in order to describe stroke focality and severity and predict rehabilitation potential (Puig et al., 2013; Puig et al., 2020). In TBI patients, a consistent reduction in fractional anisotropy (FA) has been typically found in areas affected by traumatic axonal injury (TAI). These regions include the subcortical WM of the frontal and temporal regions, the splenium of the corpus callosum, the posterior limb of the internal capsule, and the cerebral peduncles (Nakayama et al., 2006; Benson et al., 2007; Xu et al., 2007; Greenberg et al., 2008).

Combining transcranial magnetic stimulation (TMS) with electroencephalogram (EEG) is extensively used to study and assess cerebral reactivity and connectivity (Rossini et al., 2015; Ferreri et al., 2017; Tremblay et al., 2019). TMS is a non-invasive brain stimulation method that allows the study of human cortical function in vivo (Ilmoniemi et al., 1999; Hallett, 2007). TMS enables the modulation and exploration of brain functional status. Studies integrating TMS with EEG (TMS-EEG) have shown that TMS produces waves of activity that reverberate throughout the cortex and that are reproducible and reliable (Casarotto et al., 2010; Farzan et al., 2010; Kerwin et al., 2018), thus providing direct information about cortical excitability and connectivity with excellent temporal resolution (Ilmoniemi et al., 1997; Shafi et al., 2012; Chung et al., 2015; Pennisi et al., 2016; Bordet et al., 2017). TMS-EEG has been used to causally probe the dynamic effective connectivity of human brain networks (Rogasch and Fitzgerald, 2013; Ferreri et al., 2014; Kugiumtzis and Kimiskidis, 2015).

Direct electrophysiological imaging (DELPHI) is a new clinical methodology for evaluating fundamental physiological properties of brain network function by automatically combining TMS stimulation and EEG monitoring (Zifman et al., 2019). The DELPHI software algorithm extracts direct stimulation-related properties of brain network function, characterizing a profile of brain functional pathophysiology including properties of network integrity and plasticity.

Several studies have explored the correlation of TMS-evoked response to WM fibers and GM changes, indicating that this method may be useful for probing neurological disease and conditions that manifest changes in white or GM ROI’s (Kirton et al., 2010; Lanza et al., 2013; Pennisi et al., 2015).

The current study tests DELPHI technology and output parameters as a clinical tool for evaluation of specific neuronal tracks connectivity and displays its strong correlation to structural changes in WM fibers. The study population includes stroke and TBI patients that represent the population with connectivity disruption caused predominantly by WM changes and healthy controls.

The current study focuses on TMS-EEG neurophysiological insight into brain network function and connectivity by correlating TMS-evoked EEG response to WM fibers integrity. For this reason, a heterogenic group of subjects who sustained structural connectivity injuries in potentially different areas were randomly selected. The study groups include post-rehabilitation, ischemic stroke subjects with lesions of different sizes, post-rehabilitation TBI subjects, and healthy subjects. Tables 1A,B, 2A,B summarize age, gender, hand dominance, and general brain region of injury.

In the present study, we evaluated the performance of DELPHI output parameters as biomarkers for structural and functional brain connectivity abnormalities, captured by WM tracks integrity measures, in stroke and TBI patients. Study results demonstrate the changes between HC and TBI or stroke populations for DELPHI output measures. The most significant difference of both stroke and TBI patients from healthy age-matched controls was demonstrated by WFA that represents the general waveform of single pulse TEP response. Both EPD and LPD were most significantly changed between the HC and TBI/stroke groups in parietal and temporal hotspots, most probably due the higher SNR in the hotspots that are proximal to the stimulation area or contralateral to it. LPD was significantly changed in the stroke group alone, whereas EPD was significantly different in TBI alone. This may originate from the differences in age groups, as the TBI group and control are significantly younger, or might possibly arise from different natures of injuries that for stroke might be more focal and for TBI more diffused. These differences should be further explored and may potentially provide additional insight. The fact that WFA alone was significantly affected in both the TBI and stroke groups for all recording hotspots may indicate the notion that different injuries (location, sizes, etc.) are represented in different phases or latencies (late/early) of the TEP response and therefore were not as significantly noticeable in the EPD and LPD but rather in the WFA parameter that represents the entire response pattern capturing all possible changed parameters of the signal. DTI analysis demonstrated that, as expected, WM integrity was significantly compromised in both the TBI and stroke groups compared with HC in most ROIs. This is most probably due to the variability in injury locations and sizes. A multidimensional approach utilizing linear SVM model, which incorporates all DELPHI output measures, displays high classification performance of healthy vs. TBI or stroke populations with a balanced accuracy of 0.81 ± 0.02 and AUC of 0.88 ± 0.01. Moreover, a multivariate linear regression model revealed high correlation between DELPHI output measures and various WM fibers, which is consistent with the decrease of FA values in a wide range of ROIs in the stroke and TBI groups, compared with HC. Interestingly, higher sensitivity to WM integrity was observed in fibers that are in closest proximity to TMS stimulation site (M1L), namely, the highest values were recorded in the frontal and left fibers, which are proximal to the TMS stimulation site. These data support DELPHI (TMS-EEG) as a tool for neurophysiological evaluation of network connectivity and its potential as an affordable, safe and available tool for monitoring and evaluating brain network structural and functional abnormalities in specific neuronal circuits. These results are supported by previous studies (Manganotti et al., 2015; Lanza et al., 2017; Hordacre et al., 2019) and indicate that in order to get a more comprehensive and localized evaluation of functional and structural brain connectivity, integrity, and plasticity, at least two contralateral stimulation sites are required. Moreover, direct TMS stimulation of frontal, medial, and posterior areas may increase TMS-EEG sensitivity to both structural and functional ROIs. TMS-EEG technology has been shown to provide insight into the evaluation and monitoring of functional effective connectivity in various brain disorders, such as Alzheimer’s disease (AD), stroke, and TBI prognosis and rehabilitation prediction (Bortoletto et al., 2015; Manganotti et al., 2015; Tremblay et al., 2019). Our study supports the clinical utility of TMS-EEG in brain disorders of functional and structural effective connectivity, displays high sensitivity of DELPHI neurophysiological measures to WM-related structural connectivity changes as measured with DTI, and supports specific fiber localization correlated to the stimulation area. It is important to note that the current study classifiers did not consider age-related WM and GM changes, such as minor lacunar infracts, lesions, and brain atrophy, that may affect “brain health” status in the HC group; however, these measures were addressed in the regression model with FA analysis of WM fibers for all study groups. Further studies are required to address the functional sensitivity of changes in neuronal circuits and explore the potential predictive value of this tool for the early detection, predictive recovery, and prognosis of WM-related brain abnormalities and diseases. Future studies should include populations suffering from brain abnormalities that may or may not be detected by conventional clinical MRI/CT scans in order to establish detection of subtle WM changes in clinical conditions, such as mild TBI, cerebral small vessels disease (CSVD), etc., and record clinical measures of cognitive and physiological performance. The current study supports previous studies that display TEP as a biomarker for brain excitability and functional connectivity (Bagattini et al., 2019; Opie et al., 2019; Rossini et al., 2020). Here, we provide further evidence for the physiological relevance of probing specific networks (such as the motor network that was stimulated in the current study) with TMS-EEG technology in order to record structural or functional WM abnormalities in brain injuries or other diseases. These results demonstrate the potential clinical value of DELPHI output measures in differentiating brain-related diseases and abnormalities, providing a reliable, safe and easy-to-use tool for monitoring WM-related connectivity changes associated with brain-related abnormalities, such as brain injuries and degeneration.

The ability to facilitate between clinical observations, structural and functional brain connectivity, and network physiology is crucial in order to achieve an optimal clinical care. The data presented in this study support the concept that TMS-EEG physiological measures provide specific accessible insight into the human connectome that may help to narrow the gap between anatomical data, effective network function, and clinical observations, thus, providing a clinical tool for monitoring brain network function and brain health.

The DELPHI automated acquisition and analysis system can be used in order to monitor brain health throughout aging and may enable the early detection of abnormal pathophysiological changes leading to neurodegeneration.

The original contributions presented in the study are included in the article/supplementary materials, further inquiries can be directed to the corresponding author/s.

The studies involving human participants were reviewed and approved by Shamir Medical center review board. The patients/participants provided their written informed consent to participate in this study.

NZ with OL-L, HF, and ID have determined the study design and criteria. NZ conducted most of the laboratory works. OL-L and HF designed most of the DELPHI software algorithm. HF and ID contributed in designing and writing most of the manuscript and submitting the manuscript for publication. SE was principal investigator in the study. ES was responsible for the MRI imaging analysis in the study. DH, DT, SE, and ES contributed in assuring the methods and the quality of the results and with reviewing the manuscript and approving for publication of the content. All authors contributed to the article and approved the submitted version.

OL-L, NZ, DH, DT, HF, and ID are employed by QuantalX neuroscience. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.