Neural Plasticity in Human Brain Connectivity
Positive clinical outcomes are now well established for deep brain stimulation, but little is known about the effects of long-term deep brain stimulation on brain structural and functional connectivity.
Here, we used the rare opportunity to acquire pre- and postoperative diffusion tensor imaging in a patient undergoing deep brain stimulation in bilateral subthalamic nuclei for Parkinson’s Disease. This allowed us to analyse the differences in structural connectivity before and after deep brain stimulation. Further, a computational model of spontaneous brain activity was used to estimate the changes in functional connectivity arising from the specific changes in structural connectivity.
We found significant localised structural changes as a result of long-term deep brain stimulation. These changes were found in sensory-motor, prefrontal/limbic, and olfactory brain regions which are known to be affected in Parkinson’s Disease. The nature of these changes was an increase of nodal efficiency in most areas and a decrease of nodal efficiency in the precentral sensory-motor area. Importantly, the computational model clearly shows the impact of deep brain stimulation-induced structural alterations on functional brain changes, which is to shift the neural dynamics back towards a healthy regime. The results demonstrate that deep brain stimulation in Parkinson’s Disease leads to a topological reorganisation towards healthy bifurcation of the functional networks measured in controls, which suggests a potential neural mechanism for the alleviation of symptoms.
The findings suggest that long-term deep brain stimulation has not only restorative effects on the structural connectivity, but also affects the functional connectivity at a global level. Overall, our results support causal changes in human neural plasticity after long-term deep brain stimulation and may help to identify the underlying mechanisms of deep brain stimulation.
Anatomical connectivity networks derived from DTI data. The pre- and post-operative structural networks (left and right columns, respectively) are shown superimposed on a rendered brain (A/B) and as connectivity matrices, Cpre and Cpost (C/D). In both representations, the full 90-node networks are highlighted in green, while the left and right hemispheres are highlighted in blue and red, correspondingly.
What is the procedure about?
Deep brain stimulation (DBS) is a neurosurgical procedure that is increasingly used to alleviate the symptoms of a number of otherwise intractable disorders including Parkinson’s Disease (PD), essential tremor, dystonia and chronic pain. DBS for PD has become well established since the 1990s with two main surgical targets, namely the subthalamic nucleus (STN) and the globus pallidus internal (GPi) .
Recently, another target in the pedunculopontine nucleus (PPN) has also shown promise. The PPN is a relatively new target in treating primarily gait and posture symptoms in PD. Recent studies and reviews show positive results. Some studies however report much less positive results. Although positive clinical outcomes have now been well established, little is known about the effects of long-term stimulation on brain structure in terms of grey and white matter connectivity and the underlying neural mechanisms.
Some insight has come from a rat study using the 6-OHDA model of Parkinson’s disease showing that prolonged high frequency stimulation of the STN has a neurorestorative action. This finding suggests that DBS can change brain connectivity, corroborating previous findings from a human case study which, using diffusion tensor imaging (DTI) in PD, has demonstrated that DBS in the pedunculopontine nucleus (PPN) has a restorative action and increases connectivity of the PPN with the cerebellum.
How it affects the whole-brain connectivity?
What is not yet clear is how DBS affects whole-brain connectivity. The current study describes the case of a patient with PD undergoing bilateral STN DBS surgery. In this rare case we were able to acquire both preoperative and five-month postoperative DTI from this patient, which allowed us to investigate the long-term effects of DBS for PD using brain connectivity measures and computational modelling. This study involves a rare case due to the safety concerns and major artefacts related to postoperative MRI and DTI data acquisition.
This study aims to elucidate long-term effects of DBS on the topological properties of structural brain networks, investigate the restorative function of DBS and predict the dynamical impact of such structural alterations on resting-state functional networks.
To begin with the structural networks, using advanced graph theoretical measures we concentrate on measures of nodal efficiency. Nodal efficiency is related to shortest path length and is believed to characterize the ability of parallel information transfer and large scale functional integration in brain networks. Alterations in the nodal efficiency of structural brain networks in PD may be linked to the cognitive and behavioural problems occurring in the advanced stages of the disease.
How is the restorative process connected with healthy nodal efficiency?
Therefore, the recovery of healthy nodal efficiency values after DBS could be indicative of an effective restorative process in PD mediated by DBS. Local alterations in white-matter structural connectivity, as observed here in PD before and after DBS, can have significant impacts on the large-scale dynamics of the brain. As a matter of fact, several works have shown that the structural connectivity, revealed by DTI, strongly shapes the functional connectivity (FC) between brain areas during rest (measured as the temporal correlations of the blood-oxygen-level-dependent (BOLD) signal recorded with functional MRI (fMRI)).
However, the relationship between anatomical and functional brain connectivity is not trivial, and computational models of large-scale neural dynamics are unique tools to explore this relationship. Importantly, models can be used to predict the effects of structural alterations on brain dynamics, which is beyond reach on the experimental side, making models a unique tool for the comprehension of brain diseases.
Methods of investigation
To investigate the dynamical impact of the structural changes occurring in PD before and after DBS, we used a dynamic mean field model of spontaneous activity. The spontaneous dynamics obtained with the different structural connectomes, i.e. pre-DBS, post-DBS and healthy controls, was analysed in terms of stability and BOLD functional connectivity. In a previous work, it was found that the optimal fit of the model with healthy resting-state functional connectivity was obtained just before the bifurcation point, i.e. the point above which neural activity becomes unstable or chaotic, suggesting that the brain at rest operates at the edge of instability.
What was found in the study?
In the current work, we observed that this bifurcation point was shifted in PD.
Notably, it was found that DBS induced the recovery of the structural connectivity, so that the bifurcation point was shifted back towards healthy values. In addition, we compared the simulated BOLD functional connectivity obtained with the different connectomes with a typical FC from healthy controls,
We found that, despite the shift of the bifurcation occurring in PD, the dynamics exhibits homeostasis, i.e. the optimal fit with the empirical data was always found just below the bifurcation threshold, independently of the structural connectome considered. Finally, the topological properties of simulated functional networks were analysed using measures from graph theory. Results indicate that DBS induces a topological reorganisation of functional connectivity, recovering its properties towards healthy values.
Nodal efficiency changes between pre- and post-DBS structural networks. The AAL regions with more than 20% difference in nodal efficiency between pre- and postoperative measures are plotted on three-dimensional renderings of the human brain in MNI space seen from above (top) and from the side (bottom).
We predicted that long-term DBS concomitant with alleviation of some clinical symptoms would affect regions which are known to be connected with the STN and which change in PD. As such, we predicted changes in the frontal cortices, which have been shown to have high coherence with the STN in PD. We also predicted changes in the olfactory system given that olfactory dysfunction is a common symptom in PD. Prevalence estimates of olfactory dysfunction in PD vary between 70–90% of patients.
How is the olfactory system relevant to the PD?
Yet, the importance of the olfactory system in PD lies not only in its high prevalence. Olfactory dysfunction could be the first sign of PD and appears approximately five years prior to the onset of any motor symptoms. Hummel and colleagues have found that DBS of the STN in patients with PD significantly improves odour discrimination when DBS is turned on.
No effect however was found for odour detection threshold, indicating changes in higher order olfactory areas. Given these previous behavioural results investigating the effects of STN DBS on olfactory functioning in PD, we predicted a change in nodal efficiency in olfactory regions including the primary olfactory cortex as well as in the secondary olfactory cortices in the orbitofrontal cortex. Finally, we predicted that the functional connectivity resulting from the computational modelling of the brain networks post-DBS would show more similar results to that of a healthy functional connectivity compared to the strongly PD networks in pre-DBS state.
Notes from the editor
This article was reproduced under the Open-access Creative Commons Lincence from PLoS One.
Copyright (C) 2014 van Hartevelt et al.
Tim J. van Hartevelt, Joana Cabral , Gustavo Deco, Arne Møller , Alexander L. Green , Tipu Z. Aziz, Morten L. Kringelbach
The headings were introduced by the editor and are not the authors’ work. The full article can be found in PLoS One Open Access. The authorship highlighted on this blog is given to the first author of the work, with biography supplied by Department of Psychiatry, University of Oxford.