Neuromodulation for Cognitive Disorders: In Search of Lazarus?
Keywords: Alzheimer's disease, deep brain stimulation, neuromodulation, noninvasive
Cognitive decline and in particular dementia are major causes of disability in the US and the world. Alzheimer's disease (AD) is the most common type of dementia likely affecting 70-80% of this patient population., It is estimated that currently 24.3 million patients suffer from AD worldwide, and this number is projected to increase to 81 million by 2040. The disorder places significant emotional and financial toll on patients and their care givers. Many care givers are older themselves, or care both for an aging parent and for children under 18 at the same time. The societal economic burden is large. The cost of care for those with dementia in the US is estimated to be 305 billion annually, including nearly 206 billion in costs to Medicare and Medicaid, and is projected to increase to 1.1 trillion by 2050 as the population ages. 244 Billion worth of care is estimated to be provided by family member and other unpaid caregivers annually.
Despite this large footprint, pharmacological treatments for AD and other types of dementias have not been very effective and as such there has been intense interest to develop other treatment modalities with better efficacy.
Neuromodulation has proven efficacious in some neurological disorders. Deep brain stimulation (DBS), the most widely used method of neuromodulation has become an accepted treatment for Parkinson's disease when patients develop fluctuating responses to medications or side effects from high doses of medications. More recently, DBS has been FDA approved for certain epilepsy disorders and for obsessive-compulsive disorder, and there have been attempts to utilize it for depression, Tourette's disorder and others. Efforts at demonstrating efficacy for AD have also been pursued.
More recently other potentially promising noninvasive neuromodulatory treatments have been developed parallel to the DBS efforts, albeit without yet sufficient evidence of efficacy in humans.
In this manuscript, the current understanding of the pathophysiology of dementia, and in particular AD, is reviewed. We also explore recent efforts at invasive and noninvasive neuromodulation for the treatment of these disorders.
Memory encoding and access has been attributed to the Papez circuit. This complex circuit connects several components of the limbic system. Neurons from the hippocampus project to the septal nuclei and the ventral striatum via the fornices which divide at the anterior commissure to also project to the mamillary bodies., The mammillothalamic tract projects from the mammillary bodies to the anterior thalamic nuclei. These nuclei then project via the cingulum to the entorhinal cortex. The circuit is completed as neurons from the entorhinal cortex project to the dentate gyrus of the hippocampus and the subiculum via the perforant and direct pathways.,
Memory formation is thought in part to be mediated by specific firing patterns in this circuitry. Populations of neurons generate activity that may oscillate at different frequencies. Theta (3–8 Hz) and Gamma (30–100 Hz) frequency bands are thought to be prominent in the hippocampus and to play an important role in memory. Synchronization of these oscillations is postulated to serve as an important role in communication and information transfer in the connected regions of the memory circuitry. Gamma phase synchronization is deemed important in neural communication and long term potentiation, and has been implicated in memory encoding and retrieval between the entorhinal cortex and hippocampus.
Types of memory
Memory is generally categorized as short and long term. Short term memory involves handling information for cognitive tasks. It is related to, and overlaps with, working memory.
Long term memory is thought of as implicit and explicit memory based on retrieval patterns. Implicit memory, activated by associations between environmental cues and responses, is involved in remembering information unconsciously and effortlessly. Explicit memory is the part of memory that requires an active effort for retrieval. Types of explicit memory are episodic, which can be thought of as a chronological ongoing memory of the person, and semantic, which involves general knowledge about the world including facts and ideas.
Each type of memory uses subsets of the main memory circuitry. The anterior nucleus of the thalamus and the hippocampus for example, are major components required for episodic memory, as well as semantic memory. The entorhinal cortex and the parahippocampal gyrus on the other hand, are important in semantic memory.
AD disease pathophysiology
The deposition and retrieval of memory is dependent on the proper functioning and interaction of the complex components which include the circuit of Papez. The normal activity of these components in turn is dependent on the health of the neuronal components which are affected in AD. The cellular mechanisms that lead to cognitive decline in AD are complex. The “amyloid cascade theory” is a widely referenced process which describes the steps in development of AD as the deposition of extracellular depositions of amyloid beta (Aβ) peptides as neuritic plaques, and intracellular phosphorylated tau proteins as neurofibrillary tangles.,, These toxic oligomers promote the development of AD leading to cell dysfunction and death by affecting microtubule, mitochondrial, and synaptic function. It is believed that the pathophysiology of AD develops in the lateral entorhinal cortex and then spreads to other areas.,
Boncristiani and others, demonstrated in animal models and cultures that increased in Aβ deposition in cortical regions can result in reduced acetylcholine (ACh) activity. The cholinergic hypothesis by Bartus describes AD as a consequence of reduced ACh production.
The increase and reduction of ACh has been demonstrated to correlate with memory capacity in both demented and healthy patients.,, It also appears that the severity of symptoms is dependent on the degree of depletion of ACh.,,
The reduced ACh activity is then thought to be contributing to aberrant network oscillations which are thought to contribute to cognitive decline in AD.
As the disease progresses, there appears to be a disruption of the default mode network (DMN).,, The interplay of default mode network activity and gamma and theta oscillations may be an underlying mechanism of learning during tasks.
The main source of ACh in the brain are neurons in the basal forebrain. These neurons project to all layers of neocortex and the hippocampus. The nucleus basalis of Meynert (NBM) is the largest of group of nuclei in this region. The importance of this area in the pathogenesis of AD is highlighted by some more recent studies. Studies have demonstrated that there is significant atrophy in the NBM of patients with mild cognitive impairment (MCI). and the degree of atrophy correlates with degree of decline. In animal studies, it has been demonstrated that lesioning of the NBM results in a decrease in ACh and increase in the deposition of Aβ in the cortex. Given the importance of the neurons in the NBM it has been postulated to address the cascade of injury occurring at the NBM, by supporting the health of neurons in the NBM. Nerve growth factors (NGF) may offer one possibility to this end. It has been demonstrated that NGF levels are low in patients with AD, and that administration of NGF to the NBM in animal models may reverse age-related brain atrophy and improve learning and memory.,
The administration of NFG and other growth factors as a treatment for dementia and AD in patients however has been limited, including for the impermeability of the blood brain barrier (BBB) to NGF.,
Neuromodulation for neurological disorders
Efforts at neuromodulation for the treatment of AD have been attractive because of limited progress with pharmacological agents for the treatment and or reversal of AD, and the relative success of treating other neurological disorders with these modalities.
Here we will review the current developing modalities for neuromodulation in the treatment of AD and dementia. There are currently both invasive and noninvasive techniques that are at differing stages of investigation with varying degrees of evidence behind them. We have included the modalities that are in some stage of clinical research.
The lack of adequate pharmacological options, the risks as well as expense of surgical clinical trials, and advancement in technologies have led to a growth of noninvasive neuromodulation modalities for the treatment of dementia and AD. These modalities are along a spectrum of both translational evidence and progress.
Dysregulation of brain default mode network (DMN) has been associated with dementia and AD. In animal models of AD, it has been demonstrated that transcranial near-infrared photobiomodulation (PBM) may restore some cognitive improvements,,,, 2015) and PD.,,
The mechanism of action in PBM is believed to involve stimulation of mitochondria by absorption of photons in cytochrome C oxidase, presumably reducing oxidative stress.,,
A human study of 5 patients with mild to moderately severe AD treated over 12 weeks with PBM with a wearable, transcranial and intranasal PBM device, demonstrated significant improvement in cognitive scoring.
Based on the promising results, as a follow-up, eight patients were randomized to 12 weeks of usual care or 12 weeks of home PBM. A cognitive inventory at baseline, 6 and 12 weeks was obtained. Perfusion MRI and resting-state fMRI at baseline and 12 weeks were obtained. The investigators reported improvement in cognitive measures, and cerebral perfusion and connectivity within the nodes of the DMN in the PBM group.
Memory consolidation occurs in deep slow sleep. Deep slow sleep facilitates the memory consolidation stage by enabling information transfer via nodes of the memory circuitry. Sleep deprivation can affect new memory consolidation. The evidence for association of sleep and memory is robust, and based on this body of evidence it is plausible that memory can be improved by promoting slow-wave sleep synchronization.
Indeed in healthy volunteers, a newly developed method of acoustic brain stimulation at night by pink noises has been shown to perhaps improve memory consolidation. Based on this, a group is France in conducting a proof of concept study to assess the effects of auditory brain stimulation on memory capacities in a group of subjects with AD.
The role gamma oscillations play in human memory was touched upon previously. It has been established that gamma oscillations facilitate memory activity as a form of interaction within the nodes of the Papez circuitry. AD patients have been shown to have reduced neural oscillations in the gamma band (30–120 hz)., In animal models of AD, restoring these oscillations has been shown to reduce Aβ plaques and phosphorylated tau proteins which are thought to be toxic deposits in the brain, and to reduce neurodegeneration. This in turn has demonstrated improvements in memory in the animal models. The ability to noninvasively induce gamma entrainment with gamma-band light stimulation has also been demonstrated.
Researchers at MIT have developed a noninvasive device which uses visual and acoustic stimuli and can induce gamma oscillations in the brain. There is an ongoing trial to assess the effect of gamma entrainment, or induction of gamma oscillation, in patients with mild AD via this device. The investigators are looking to recruit 40 patients with mild AD for assessment of feasibility, tolerability and safety of gamma frequency entrainment in this population.
Magnetic and electromagnetic neuromodulation
Magnetic stimulation has evolved as an alternate way to deliver electrical stimulation to the brain (versus ECT) as a treatment for certain disorders like depression. While the mechanism of effect is not greatly understood, there is some consensus that cortical activity can be modified by repetitive delivery of a magnetic field. One possible theory for the effect of rTMS on memory is that rTMS may facilitate LTP-like changes in synaptic connectivity. LTP may regulate the expression of neurotrophic factors which decline in AD, and animal studies confirm that high-frequency rTMS increase some trophic factors. Repetitive Transcranial Magnetic stimulation (rTMS) has been used for treatment of mood disorders and was FDA approved for the treatment of depression in the US in 2008. In 2010 a large randomized study demonstrated efficacy of rTMS versus sham in patients with depression to maintain remission.
Investigators have used this modality for treatment of other neurological disorders and in particular several small trials have been undertaken in AD and cognitive decline.
Transcranial electromagnetic treatment (TEMT)
In contrast to rTMS stimulation, where magnetic waves radiate from and return to a conductor, Transcranial Electromagnetic Treatment (TEMT) uses an emitter source where magnetic and electric waves are just dispersed and do not return to the source.
In mouse models for AD or human AD cells, transcranial electromagnetic treatment has been shown to disaggregate both Aβ and phosphorylated tau oligomers,, as well as α-synuclein oligomers. As these toxic oligomers are believed to be the initiators of the AD pathophysiology, a modality that facilitates their breakdown may hold promise.
In an open-label study of eight mild to moderate AD patients, who received TEMT over a 2-month period, investigators demonstrated a reversal of cognitive impairment in key tasks, with changes in CSF levels of Aβ consistent with a disaggregation of the molecules in the brain. In the same study, functional imaging demonstrated increased neuronal activity and connectivity.
Ultrasound sonication has been demonstrated to have the potential of targeted modulation of the brain with accuracy., The effect of this modality on tissue can be either thermal or mechanical and the extent can be either temporary or permanent based on the energy delivered.,,,,, As such there has been tremendous interest in applying this modality for treatment of neurological disease. Here we discuss two separate treatment strategies and the investigational efforts.
Transcranial pulse stimulation
Investigational studies have shown that focal ultrasound applications can affect the human brain function., Transcranial pulse stimulation (TPS) uses a single short ultrasound pulse to sonicate the brain target. This modality uses no periodic waves and has better skull penetration. The mechanism of action is postulated to be a mechanical effect on cell membranes that can affect cell ion channels, and resultant changes in neurotransmitter levels and increases in growth factors., In an AD mouse model, ultrasound sonication has demonstrated microglial activation and Aβ plaque reduction.
A preclinical investigational pilot of 35 patients with AD has demonstrated that stimulation of the AD network (dorsolateral prefrontal cortex, inferior frontal cortex, bilateral lateral parietal cortex) improved memory performance with concomitant changes in functional imaging demonstrating an increase in connectivity.
Blood brain barrier opening
One of the possible reason for failure of several clinical trials of innovative drugs including antibodies to Aβ, for the treatment of AD, may be lack of permeability of the BBB to these therapeutics. Low-intensity focal ultrasound has been shown to be able to open the BBB in a noninvasive way.,,, This reversible opening time can be modulated by the intensity of the ultrasound. Targeted ultrasound has been used to increase delivery of anti Aβeta antibodies to transgenic mice, and this has demonstrated decrease in the Aβ load in the treated mice. What is interesting is that the BBB opening by ultrasound without any anti-Aβ treatment was able to also reduce the Aβ in mice in the same study. This is presumably to do the opening of the BBB allowing endogenous antibodies to attack the Aβ. Following the BBB opening, the transgenic mice demonstrated improved performance on memory tasks. Regardless of the mechanism, the prospect of potential treatment option for patients with AD with this noninvasive modality is exciting.
There are at least two devices that are utilizing this concept in the treatment of patients with AD.
SonoCloud. (CarThera, Paris France)
These investigators have developed a small device which is surgically implanted under the skull. Implantation on the skull allows the bypass of bone and thus ultrasound signals are not degraded. They have demonstrated efficacy in opening the BBB for chemotherapeutic treatment of gliomas. BOREAL 1 trial (NCT03119961) is a phase 1–2 study to assess BBB opening in AD using this technology.
Another group of researchers demonstrated the use an noninvasive MRI-guided device, Insightec (Tirat Carmel, Israel) in a phase 1 trial of 5 patients. They demonstrated that the BBB can be opened safely, reversibly and repeatedly with this technique. While they did not demonstrate any improvement in cognition or PET imaging, the demonstration of feasibility has encouraged further efforts.
There is currently a multisite phase 2 trial aiming to enroll 20 patients to assess the safety and efficacy of this technique for opening of the BBB.
Vagal nerve stimulation
In the 1880s, James Corning, a New York neurologist, believing that venous hyperemia was causative for seizures, developed a carotid compression and vagal stimulation device. These concepts did not gain mainstream acceptance until a century later when several studies in the 1980s demonstrated the reduction of seizures by stimulation of the vagus nerve. The FDA subsequently approved the use of a device for vagal nerve stimulation in 1997 for medically refractory epilepsy. The use of this device has been also investigated in other disorders, with varying success.
A study of vagal nerve stimulation in 17 patients with AD dementia showed 41.7% with improved cognition at 1 year after treatment. There does not seem have been any follow-up after these seemingly positive results however.
Deep brain stimulation (DBS) is an established surgical procedure for the treatment of Parkinson's disease, essential tremor and dystonia. There are also more recent indications for use in obsessive-compulsive disorder and epilepsy. And there are ongoing efforts to demonstrate its efficacy in depression and other disorders as well.
Essentially a “pacemaker” for the brain, the system consists of wires or “leads” implanted in specific targets within the brain. The leads are then connected to a battery or “pacemaker,” which is then programmed by the clinicians to deliver an electrical stimulus to the brain.
Theoretical effects of DBS
There is substantial research conducted on DBS, and while there is no general consensus on the exact mechanism of this modality, there are some more widely cited theories. There is evidence that local stimulation may inhibit cell bodies and at the same time stimulate axons near the field of stimulation., DBS may also be having effects on brain-wide networks. In animal models of thalamic stimulation, changes in hippocampal neurotransmitter release have been demonstrated, indicating modulation of distant targets.
And yet another potential possibility may be the effect of DBS on neuronal oscillatory activity., DBS also may be affecting memory by affecting neuronal oscillations which are thought to be methods of communication between clusters of neurons, and responsible for some forms of memory formation. DBS may be able to modify pathological forms of oscillation,, allowing more stable forms of oscillation to emanate.
It is also possible that DBS may be facilitating neurogenesis. In a study where neurogenesis in rats was presumably reduced by administration of corticosteroids, introduction of stimulation resulted in improved cognitive performance, but only if there was a delay of over 30 days between the stimulation and testing presumably this time after stimulation needed for neurogenesis.
Another potential mechanism of action of DBS may be the release of growth factors by electrical stimulation.,, Studies in cell cultures have shown the release of NGF following stimulation and this has been also demonstrated in animals where unilateral NBM stimulation resulted in increase in ipsilateral extracellular NGF.
Several targets for DBS have been pursued for modulation of cognition in patients with AD based on laboratory and preclinical body of evidence. In the following section several of these targets will be reviewed with a brief description of supporting evidence.
Forniceal stimulation in animal models has been shown to improve memory. The effect of DBS may be stimulation of large myelinated axons of the fornix., It appears in turn that this stimulation may upregulate growth factor expression, which may in turn lead to changes in long-term potentiation.
While treating a patient with morbid obesity with ventromedial and lateral hypothalamus DBS, vivid memories of past events were elicited when the stimulation field was near the fornix. The subject had improved scoring in cognitive testing and functional imaging.
Based on this, a phase I trial of forniceal stimulation in humans was conducted. 6 patient with mild AD underwent bilateral fornix DBS, and were assessed with cognitive and imaging follow-up for 1 year. The results demonstrated an overall lower rate of decline with increases in glucose metabolism. These positive results lead to the expansion to a phase II study.
ADvance I study. Phase II study of fornix DBS stimulation in mild AD
This multicenter, double-blinded, randomized controlled trial, recruited 42 patients with mild to moderate AD who underwent bilateral fornix DBS. The patients then received sham versus active stimulation for 12 months. The primary endpoint was cognitive testing at 12 months. Interestingly, the study did not demonstrate any improvement in cognitive status of the patients who received stimulation compared to controlled. In a post hoc analysis; however, patients over 65 showed less deterioration in their cognitive scores.
Glucose metabolism in patients over 65 also increased by 14-20% at 12 months. In summary while there were no differences in primary end points, there may have been some benefit for patients over 65.
ADvance II study Phase IIb/III
Investigators initiated a larger follow-up study aiming to more definitively test the efficacy of DBS in AD, and guided by the post hoc analysis of the ADvance I study. This phase IIb/III trial is currently ongoing and the goal is to recruit up to 148 patients with AD over 65. The primary end point of the study is changes in the integrated AD disease rating scale (iADRS). Additional assessments of the patients will include glucose PET imaging and CSF studies.
Nucleus basalis of Meynert
The NBM appears very involved in the pathogenesis of dementia in AD and Parkinson's disease (PD). 96% of the neurons in this region have been demonstrated to be lost at autopsies of AD patients compared to same age nondementia controls and this has also been shown in imaging of AD and PD dementia patients.
The importance of this cholinergic pathway has been the impetus behind consideration of DBS of this nucleus. The theory is based on the connectivity of this region to the medial temporal structures, the amygdala, and the temporoparietal and frontotemporal association areas. Stimulation of the NBM is thought to be able to increase cholinergic activity in the association areas, and perhaps improve memory and cognition. Indeed animal evidence shows that low-frequency stimulation of the NBM at 20-50 hz can result in ACh release in the cortex, and NBM DBS in rats has been shown to enhance memory retention and acquisition.,
The first attempt at NBM DBS in a human was in 1984, where a DBS lead was placed unilaterally in the NBM. Low-frequency stimulation was delivered with cycling between 30 seconds on and 12 minutes off. While no clinical improvement was observed, stabilization of functional imaging was demonstrated on the stimulated side.
Phase I study
Based on the aforementioned evidence, a phase I clinical trial was undertaken to determine safety of NBM DBS stimulation. 6 patients with mild to moderate AD underwent bilateral NBM implants. The patients were randomized to 2 weeks of low-frequency stimulation followed by sham stimulation. After the initial randomization, all patients received active stimulation for the open-label arm for 11 months. The primary endpoint in the study was changes in cognitive scoring at 12 months. The results demonstrated some less than expected deterioration. There currently are no active large NBM DBS trials.
Hippocampus and Entorhinal cortex
The results of the hippocampal and entorhinal stimulation on memory have been somewhat contradictory.
In a multicenter study of 49 epileptic patients who had DBS stimulation in hippocampal and entorhinal areas, spatial and verbal memory encoding was noted to be significantly worse in patients that had stimulation.
On the other hand, another study of 7 epileptic patients found that entorhinal stimulation with DBS improved spatial memory performance. The potential contrary results may have to do with the type of memory task assessed in the two studies, as well as the duration of stimulation in each trial.
A potential mechanism of improved memory with hippocampal and entorhinal stimulation is alterations in network oscillatory activity which may be more conducive to long-term potentiation. Which in turn may be facilitating neurogenesis. Currently, there are not ongoing large clinical trials for hippocampal/entorhinal DBS for AD.
Anterior thalamic nucleus
The evidence for anterior thalamic stimulation (ANT) for cognitive disorders is also seemingly contradictory. Animal studies with ANT DBS stimulation have demonstrated impaired memory. Other researchers have found that high-frequency stimulation of the ANT may result in neurogenesis in the hippocampus. In the SANTE trial (Stimulation of the Anterior Nucleus of the Thalamus for Epilepsy) 10 patients with epilepsy underwent bilateral ANT DBS for assessment of efficacy for epilepsy. As a secondary measure, cognition and mood were assessed in these patients post-stimulation. While there was no statistical difference in cognition and mood, more depression and memory problems were reported as adverse effects., Currently there are no ongoing investigational efforts for ANT DBS for cognitive disorders.
For the past two decades, many promising pharmacological options for dementia and in particular AD have met lukewarm results. With the ever-growing population of elderly and the projected increase in the number of people suffering from various forms of dementia, there is a clear need for more effective therapeutics. Novel technologies and re-evaluation of established technologies based on a varying background of preclinical evidence have rekindled hope for new therapeutic options for patients suffering from cognitive decline and dementia.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.