Towards Next-Generation Detection and Modulation of Pathological Network Activity

Sameer A. Sheth
Guy McKhann II
Chuck Mikell

FUNCTIONAL NEUROSURGERY IN 2065

Of the neurosurgical subspecialties, functional neurosurgery is the one that has the greatest potential to become emblematic of scientific progress of the age. Just as the moon landing signified advances in aeronautics in the 1960s, a paralyzed child able to control a wheelchair with a cortical implant may signify the progress of the last decade.1 In addition to spinal cord injury, restorative neurosurgical therapies are being investigated for a number of devastating diseases, including traumatic brain injury,2 autism,3 and Alzheimer’s disease.4 Although our goals are lofty, our current interventional tools are relatively crude. Our electrical stimulation techniques, which have not substantially changed in 30 years, target brain regions but not cell types, and typically lack feedback control. Non-invasive stimulation methods using magnetic or electric fields have limited specificity, especially for deep structures. Functional genetic modulation is still in its infancy. In recognition of this technological gap, development efforts are underway on truly novel interventional strategies, including nanowire electrodes5 and optical interrogation of brain networks.6 While these technologies hold promise, we suggest that the next generation of treatments for functional disorders should require modulation of global brain networks with high spatiotemporal resolution and broad coverage (Figure 1).

Here, we make the case that the arc of neuroscientific progress has been in the direction of increased spatial and temporal precision, and we explore how this work has illuminated our understanding of how brain circuits support behaviors.

Figure 1: Advances in neuroscience are constrained by spatial and temporal resolution  and by locality of sampling. In 1965, postmortem histological and histochemical  techniques were available, which had high spatial precision along with global sampling, but essentially no temporal resolution (left panel). Electroencephalography (EEG) allowed  for some spatial and temporal resolution, with fairly wide sampling. In 2015, current techniques represent tradeoffs, exemplified by that between fMRI (high spatial resolution, low temporal resolution, global sampling) and single cell recordings (high spatial and temporal resolution, extremely local sampling) (center panel). DBS is essentially a local technique. In 2065, the putative technology may allow for interrogation of brain networks  with high spatial and temporal precision, along with global sampling (right panel).  Moreover, intervention will hopefully share these characteristics.

We discuss how constraints on resolution and coverage have led to the currently available models of the etiology of brain disorders. We then suggest that current terminology (such as “bipolar disorder” or “seizure disorder”) can hinder progress by imposing artificial diagnostic categories on highly idiosyncratic functional abnormalities that vary tremendously between individual patients. Finally, we speculate on the tools that will be needed to interrogate and modulate brain networks with the required precision and coverage, an approach that will truly individualize diagnosis and treatment.

As neurosurgeons, we are uniquely positioned to contribute to this technological development by virtue of our access to the brain and our preparedness to capitalize on neuroscience developments to create novel treatments.

Neurosurgeons have a strong track record of translating basic science findings into the clinic, the ICU, and the operating room. Although this review is necessarily speculative, we believe some general themes can be extracted from progress made in previous decades and that these themes can be used to conjecture about where the field is headed.

Advances over the last 50 years have been driven by progress within neurosurgery as well as the broader realms of medicine, psychology, and neuroscience. The etiology of Parkinson’s disease is an illustrative example. In the 1950s, it was observed that the blood pressure agent reserpine caused exacerbation of Parkinsonian symptoms by depletion of dopamine;7 around the same time, histochemical techniques subsequently confirmed that the striatum contained most of the dopamine in the mammalian brain. A crude model of the “extrapyramidal” dopamine system was proposed, in which excess dopamine in the striatum led to hyperkinetic disorders, and decreased dopamine led to Parkinson’s disease.8 The common theme of these observations was their very low spatial and temporal resolution.

A key prediction of this model was soon realized: Walther Birkmayer administered intravenous levodopa to Parkinson’s patients, leading to marked improvement of symptoms, and the first rationally designed neurobiological therapy was born.9 Despite this major advance, the explanatory power of this model was limited, and it lacked the ability to make sophisticated predictions about how dysfunction in dopamine circuits leads to movement dysfunction.

In recent years, neuroimaging and primate single-neuron neurophysiology have enabled more sophisticated interrogation of the dopamine system. These techniques offer major advantages over previous biochemical and histological techniques. Neuroimaging has very high spatial resolution and generally global sampling; however, it has limited simultaneous temporal resolution. Similarly, single cell recordings have high temporal and high spatial resolution but very limited coverage. Nonetheless, these techniques allowed for the development of a more sophisticated model of dopaminergic transmission. In this view, dopamine transmission is segregated into “direct” and “indirect” pathways through the striatum, which promote and inhibit movement, respectively.10

Deep brain stimulation of the subthalamic nucleus grew out of a conscious attempt to inhibit the indirect pathway.11, 12 However, despite its efficacy, DBS is ultimately limited in its effects; it has moderately high spatial precision (brain regions but not cell types) but only the grossest temporal resolution (i.e., the ability to be turned on or off). Although it represents a major advance in the treatment of Parkinson patients, these limitations may be why DBS has thus far had mixed results in the treatment of behavioral disorders.13

Behavioral disorders (including neuropsychiatric and developmental disorders) account for a considerable portion of human suffering. Because of their prevalence, they likely represent a major component of the future of functional neurosurgery. For several reasons, however, they have been very difficult to understand in neurobiological terms. First, the need for diagnostic terminology has led to inappropriate “silos” between disorders that mean little to individual patients. For instance, obsessive-compulsive disorder and depression are comorbid in approximately a third of patients.14 While considerable research has gone into each disease, the fact that so many patients meet criteria for both diseases suggests that the distinction is more fluid than the terminology supposes.

Alternative frameworks have been proposed, most prominently the Research Domain Criteria (RDoC), which proposes diagnosing psychiatric disease on a “matrix” of symptom dimensions, agnostic to current Diagnostic and Statistical Manual-V (DSM-V) terminology.15 It is hoped that these symptom dimensions will map more easily onto functions known to be supported by individual circuits (e.g., reward valence by the dopamine system). However, even these symptom domains are limited by terminology that may or may not be appropriate, and reflect an understanding of circuits limited by current techniques. For example, the symptom domain called “attention” is likely subserved by numerous circuits that have selectivity for various categories of stimuli, including faces, emotions, environmental cues, language, and so on. As techniques advance for studying the nervous system, these categories may require refocusing.

In the neuroscience of the future, it seems likely that global or brain-wide networks will be interrogated with techniques that marry high spatial and temporal resolution. They would allow for single-subject analysis of how pathological thoughts or behaviors arise from corticothalamic networks (e.g., how a schizophrenia patient comes to believe that the television is sending him messages). One current, primitive view is that in patients with schizophrenia, inappropriate dopamine release leads to aberrant salience of sensory input, leading to complex delusion formation.16 The hoped-for technology would be able to identify the exact brain circuits that attribute salience in the individual patient (high spatial precision) and know when they are working well and when they are misbehaving (high temporal precision). They would not be limited to local sampling of, for instance, the ventral hippocampus, but would have global sampling ability and the ability to identify pathologic circuits from primary sensory cortex through association areas and back to motor output.

In this putative framework, the neurosurgeon of the future would be able to identify and modulate these circuits. His or her techniques would not be limited to local delivery of electrical current or genetic material; rather, the firing rates of large groups of neurons would be able to be modulated with high temporal and spatial accuracy. Healthy percepts would be allowed through, and pathologic ones silenced or repaired. One could imagine a stay in an inpatient unit akin to an epilepsy monitoring unit (EMU), where both behavior and neural activity could be monitored. To extend the epilepsy metaphor, the patient could then undergo definitive treatment when his or her individual pathology is understood, much as an epilepsy patient whose seizures localize to the hippocampus undergoes resection of the temporal lobe. It is obviously hoped that this approach would share the high success rate of epilepsy surgery in this setting.

For that matter, it seems likely that epilepsy will be treated with methods that also exploit high-fidelity techniques. It may be the case that the activity captured by conventional EMU techniques is largely post-synaptic and does not reflect underlying spiking dynamics.17 Thus, new techniques should be able to identify areas of high neuronal activity with high precision and treat them with methods that might even be noninvasive. As another example, prostheses will likely be much more effective and “human” if they can harness combined information streams from sensory, integrative, and executive networks, not just motor.

While it is difficult to conceive what direction novel therapeutics may take, the future seems to have already arrived with the advent of focused ultrasound18 and optogenetics in primates.19 What these seemingly disparate techniques have in common is the ability to modulate neuronal spiking with high spatiotemporal resolution. What they lack is the ability to do so across brain-wide networks, rather than just in certain nodes within the network. Accomplishing the proposed goal requires either the targeted delivery of energy, whether mechanical (ultrasound) or electromagnetic (electrical, magnetic, optical), or else targeted biological transformation (using exogenous genetic material or the body’s own tissues). Which of these methods prevail, or whether entirely new approaches emerge, remains to be seen.

>IN THE NEUROSCIENCE OF THE FUTURE, IT SEEMS LIKELY THAT GLOBAL OR BRAIN-WIDE NETWORKS WILL BE INTERROGATED WITH TECHNIQUES THAT MARRY HIGH SPATIAL AND TEMPORAL RESOLUTION. THEY WOULD ALLOW FOR SINGLE-SUBJECT ANALYSIS OF HOW PATHOLOGICAL THOUGHTS OR BEHAVIORS ARISE FROM CORTICOTHALAMIC NETWORKS.<

References

  1. Tseng PH, Rajangam S, Yin A, et al. Wheelchair navigation with wirelessly recorded cortical ensembles [abstract]. Society for Neuroscience, 2014.
  2. Schiff ND, Giacino JT, Kalmar K, et al. Behavioural improvements with thalamic stimulation after severe traumatic brain injury. Nature. 2007;448(7153):600-3.
  3. Sturm V, Fricke O, Bührle CP, et al. DBS in the basolateral amygdala improves symptoms of autism and related self-injurious behavior: a case report and hypothesis on the pathogenesis of the disorder. Front Hum Neurosci. 2012;6:341.
  4. Kuhn J, Hardenacke K, Lenartz D, et al. Deep brain stimulation of the nucleus basalis of Meynert in Alzheimer’s dementia. Mol Psychiatry. 2014.
  5. Kang M, Jung S, Zhang H, et al. Subcellular neural probes from single-crystal gold nanowires. ACS Nano. 2014;8(8):8182-9.
  6. Piper SK, Krueger A, Koch SP, et al. A wearable multichannel fNIRS system for brain imaging in freely moving subjects. Neuroimage. 2014;85 Pt 1:64-71.
  7. Carlsson A. Evidence for a role of dopamine in extrapyramidal functions. Acta Neuroveg (Wien). 1964;26:484-93.
  8. Barbeau A. Dopamine and basal ganglia diseases. Arch Neurol. 1961;4:97-102.
  9. Birkmayer W, Hornykiewicz O. [The L-3,4- dioxyphenylalanine (DOPA)-effect in Parkinsonakinesia]. Wien Klin Wochenschr. 1961;73:787-8.
  10. DeLong MR. Primate models of movement disorders of basal ganglia origin. Trends Neurosci. 1990;13(7):281-5.
  11. Benazzouz A, Gross C, Féger J, Boraud T, Bioulac B. Reversal of rigidity and improvement in motor performance by subthalamic high-frequency stimulation in MPTP-treated monkeys. Eur J Neurosci. 1993;5(4):382-9.
  12. Limousin P, Krack P, Pollack P. Electrical stimulation of the subthalamic nucleus in advanced Parkinson’s disease. N Engl J Med. 1998;339(16):1105-11.
  13. Morishita T, Fayad SM, Higuchi MA, Nestor KA, Foote KD. Deep brain stimulation for treatment-resistant depression: systematic review of clinical outcomes. Neurotherapeutics. 2014;11(3):475-84.
  14. Overbeek T, Schruers K, Vermetten E, Griez E. Comorbidity of obsessive-compulsive disorder and depression: prevalence, symptom severity, and treatment effect. J Clin Psychiatry. 2002;63(12): 1106-12.
  15. Insel T, Cuthbert B, Garvey M, et al. Research domain criteria (RDoC): toward a new classification framework for research on mental disorders. Am J Psychiatry. 2010;167(7):748-51.
  16. Kapur S. Psychosis as a state of aberrant salience: a framework linking biology, phenomenology, and pharmacology in schizophrenia. Am J Psychiatry. 2003;160(1):13-23.
  17. Schevon CA, Weiss SA, McKhann G Jr, et al. Evidence of an inhibitory restraint of seizure activity in humans. Nat Commun. 2012;3:1060.
  18. Elias WJ, Huss D, Voss T, et al. A pilot study of focused ultrasound thalamotomy for essential tremor. N Engl J Med. 2013;369(7):640-8.
  19. Diester I, Kaufman MT, Mogri M, et al. An optogenetic toolbox designed for primates. Nat Neurosci. 2011;14(3):387-97.