Deep Brain Stimulation of the Fornix

Alzheimer disease (AD) is associated with several key molecular pathway disturbances that contribute directly to its pathognomonic features: neurotoxic peptide accumulation, neuroinflammatory damage, and neuro-oxidative dysregulation.1-3 These pathological changes converge to produce widespread synaptic disruption in vulnerable regions of the brain in affected individuals, notably areas important for memory. Recently, deep brain stimulation (DBS) of the fornix, a white matter tract important for limbic function and episodic memory, has been trialed in patients suffering from AD.4-6 The results of these early studies suggest that forniceal DBS (f-DBS) engages circuits and networks in brain regions of individuals with AD that are potentially therapeutic, despite unyielding molecular pathology. These findings underscore the emerging idea that AD pathology can also be viewed as a dysfunction of distinct brain circuits, similar to other circuit disorders treated with DBS such as the corticobasal ganglia circuit in patients with movement disorders. This idea, backed by observations in experimental animals, has generated considerable interest in exploring the therapeutic potential of neuromodulation in patients with AD.7-36 Moreover, a phase II trial of f-DBS with a larger cohort is near completion. In this mini-review, we discuss the rationale for pursuing f-DBS as an approach for addressing memory impairment in patients with AD. We also highlight a number of unresolved topics currently being investigated with the goal of halting or reversing defects in neurotransmission and the resulting disturbance in cortical function in patients with AD. FORNICEAL STIMULATION: PREDICTED CIRCUIT EFFECTS FROM ANATOMY AND PHYSIOLOGY The fornix, a reverse C-shaped structure (Figure) that comprises the core of the medial limbic pathway, is a unique tract in the central nervous system in that it is composed of all 3 principal white matter fiber types, consisting of association, commissural, and projection elements. It directly connects medial temporal lobe structures with the medial hypothalamus, and it plays a critical role in the consolidation of emotionally salient memories that form distinct and indelible episodes in our conscious lives, ie, episodic memories. The fornix originates as a thin white matter bundle, named the fimbria, from the superomedial hippocampal surface just posterior to the hippocampal head and lateral to the choroidal fissure. Axons in the fimbria travel as the fornix posteriorly in the medial part of the floor of the temporal horn. Forniceal fibers then turn superiorly, known as the crus, and arch around the pulvinar nucleus of the thalamus in the anterior chamber of the atrium. Crossing fibers between the 2 forniceal crura create the forniceal commissure. In the lateral ventricles, the 2 bundles join into the body of the fornix and advance rostrally along the inferior edge of the septum pellucidum. As they reach the foramen of Monro, they can be seen on axial projections of brain imaging slices as separate adjacent columns of tracts that proceed inferiorly and split in an anterior-posterior orientation along the anterior commissure. The postcommissural fibers terminate in the mammillary nuclei of the mammillary bodies. The precommissural fibers terminate in the septal nuclei.27,37-41FIGURE: A preoperative magnetic resonance image in 3 planes with diffusion tensor imaging demonstrating reconstruction of forniceal fibers in a patient undergoing deep brain stimulation (inset shows postoperative magnetic resonance image). The fornix comprises the core of the medial limbic pathway and directly connects medial temporal lobe structures with the medial hypothalamus. The reconstruction was performed by identifying mammillary bodies as regions of interest with an open-source image processing software (3D Slicer, https://www.slicer.org/). Precommissural fibers are not included in this reconstruction.Lesioning studies of the fornix, principally in patients with stroke or postoperative forniceal injury, reveal specific deficits such as anterograde amnesia that suggest it is involved in explicit recall of memories.42-48 Volumetric and tractographic studies of the fornix also show correlation between its size and degree of memory impairment in individuals with AD, suggesting its potential usefulness as an imaging biomarker.49 Thus, activation of the fornix with electrical stimulation presents a theoretical rationale for use as a therapeutic target for memory modulation. Hamani et al4 noticed an unexpected trigger of autobiographical recall in an awake patient undergoing posterior hypothalamic DBS for morbid obesity. This phenomenon was replicated in 2 additional patients with AD during a phase I trial of f-DBS.5 In that same study, electroencephalography source localization identified the hippocampal formation as the earliest site of activation after the onset of stimulation. This finding is consistent with the notion that electrical stimulation of the fornix produces short-latency antidromic activation of hippocampal structures. Orthodromic Papez circuit and cortical network activation has also been shown with neurophysiologic markers and glucose metabolism neuroimaging studies of AD patients with DBS stimulation.6,31 These observations have also been complemented by a number of trophic effects seen with electrical stimulation, including local increases in neurotrophic and synaptic proteins in the hippocampus of experimental animals.50 DBS of several structures within the limbic circuit, including the fornix,36 entorhinal cortex,51 anterior nucleus of the thalamus, which receives efferents of the mammillary bodies, and the hippocampus proper, has also been shown to drive hippocampal neurogenesis in adult rodents.52-55 A second and equally important aspect of forniceal function is its role in visuospatial processing.56-65 This component of the limbic circuit has been successfully modeled in patients with intractable epilepsy undergoing temporary deep intracranial electrode monitoring and stimulation for seizure location. In a recent study, 4 epilepsy patients received stimulation of the proximal fornix near the hippocampal formation in a double-blind, randomized fashion while undergoing repeated memory function testing. The results indicated an improvement in visuospatial memory using a theta burst pattern.66 Although this study did not directly compare high-frequency continuous stimulation with burst stimulation, theta burst patterning has previously been shown to improve spatial memory performance in animal models.67 Disruptions in theta oscillations lead to spatial navigation deficits in experimental animals68 and can be reversed with artificial theta input.69,70 In prior trials of entorhinal cortex stimulation, theta phase resetting was postulated as a putative mechanism of spatial memory improvement.21,70 Hence, a goal of current f-DBS trials and animal studies is to determine whether target location within the limbic circuit (proximal vs distal), stimulation parameter dynamics (continuous vs burst), or disease model (AD vs epilepsy) influences the degree of improvement in each of these cognitive domains. Moreover, the mechanism responsible for cognitive changes seen in these trials, ie, neurogenesis, synaptic reintegration, or circuit modulation, remains unknown. FUNCTIONAL CONNECTIVITY OF NETWORKS IN AD IS ALTERED BY F-DBS An important set of mostly posterior cortical structures, but also medial prefrontal and temporal regions, known as the default mode network (DMN) is consistently found to be impaired in patients with AD. In healthy individuals, the DMN is active during mind wandering, temporary moments during wakefulness when we are cognitively disconnected from external stimuli and more focused on self-reflection and contemplation of the future.71-75 This network, called DMN because of its deactivation during active cognitive tasks, also has close correlations with other important aspects of AD, including the location of Aβ deposits and a prominent role in episodic memory.76-81 In patients with AD, attenuation of the DMN was found to be concomitant with activation of another large-scale network involving mostly frontal cortical regions known as the salience network, shown to participate principally in emotional and social processing.82 Thus, during the course of AD, network dynamics may compensate by leveraging related circuitry in other systems, and AD symptoms such as memory impairment and preserved social-emotional functioning reflect these changes.83 Patients with DBS of the fornix represent a unique opportunity to test hypotheses concerning the role of large-scale networks in AD brain activity and response to pathological insults. Although functional magnetic resonance imaging sequences are still not routine in patients with DBS electrodes, metabolic imaging such as positron emission tomography in AD patients with f-DBS shows persistent increases in glucose metabolism in parietal and temporal areas, partially recapitulating data from studies linking AD and DMN changes.5,6 Metabolic increases in these areas have been shown past 1 year in the initial cohort, which may signal that f-DBS works differently from pharmacological approaches such as acetylcholinesterase inhibitors, which usually show a return to baseline cerebral glucose metabolism levels sooner.6 ASSESSING F-DBS IN PATIENTS WITH AD: ARE WE USING THE CORRECT SCALES? Our ability to interpret direct circuit effects from f-DBS and long-term network changes in patients with AD depends to a large extent on the rating instruments we select to test memory and other cognitive functions in this population. Two tests commonly administered to establish the presence and severity of AD include the Alzheimer's Disease Assessment Scale (ADAS) and Clinical Dementia Rating (CDR). The ADAS-cog,84 a subset of cognitive domain–specific tests from the ADAS, includes scales of declarative memory (short- and long-term recall, recognition memory), praxis, language, and orientation. The CDR-SB (sum of boxes) contains a subset of cognitive tests from the CDR,85 a semistructured interview used to stage AD that relies on subjective reports of the caregiver and patient and more objective tests of cognitive function. CDR-SB is used to assess cognitive performance in the patient in the domains of episodic memory, semantic memory, orientation, judgment, and problem solving, as well as functional domains (eg, self-care, daily activities). The detection of changes in cognitive function in response to adjustment of f-DBS parameters in patients with AD would require tests that can be administered rapidly and repeatedly, preferably over a short period of time. Accordingly, the limitations of instruments such as the ADAS-cog and CDR are long administration time and lack of alternative test forms to minimize practice effects. The ADAS-cog also has shown suboptimal sensitivity to changes in AD over time.86,87 Additionally, the episodic memory tests included in these measures consist of verbal learning tasks that may be more prone to memory interference88 than pictorial stimuli, which are more memorable and distinct and have been implemented successfully in double-blind studies.89 Hence, use of pictorial stimuli may minimize interference of previously learned information with that of tests administered shortly thereafter. A potentially fruitful approach may be to develop tests that (1) are known to be sensitive and specific for hippocampal functions such as pattern separation, spatial memory, navigation, and associative learning (ie, binding)90-92; (2) allow the use of novel stimuli in each version of the task to minimize practice effects; and (3) can be administered rapidly, with automated results output for selecting optimal DBS parameters. Currently only a few tests sensitive to hippocampal function typically used in animal studies have been created93 and used in patients with AD.94,95 CONCLUSION AD originates from pathological molecular processes that ultimately result in interruptions in large-scale networks supporting memory-based cognition. The goal of f-DBS is to partially restore episodic memory formation and visuospatial processing through activation of hippocampus and mesial temporal lobe structures plus more widespread regions in the posterior cingulum and midline prefrontal cortex. Now that f-DBS has been successfully shown to be a safe and well-tolerated procedure in an initial cohort of patients with AD, future considerations of this novel approach should seek to elucidate the specific nature of stimulation-evoked changes in network function, ie, memory vs visuospatial improvements, and how those changes are achieved, ie, neurogenesis vs oscillatory dynamics. Finally, a redesign of AD rating scales customized for testing DBS parameter changes and sensitive to circuit-specific functions is needed. We are actively pursuing clinical trials to determine in which AD patients f-DBS is effective and whether stimulation of cognitive circuits with DBS could become a widely accepted therapeutic option for patients with AD. Disclosure The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.