Connectomics & Epilepsy

Epilepsy and the recurrent seizures which define the condition is one of the most disabling and unpredictable neurological conditions an individual can face. Temporal Lobe Epilepsy (TLE) is the most common form of epilepsy, afflicting over 50 million people worldwide¹. Despite its prevalence,  a complete neurobiological understanding of the condition remains elusive. Generally speaking, we understand that during epilepsy, epileptogenic foci - or, specific brain regions and networks - misfire, causing seizure onset^2,3. 

Figure 1: Epileptogenic brain networks

Since neural connections and their communication are the underlying source of epileptic events, connectomics, a technique which evaluates how independent brain regions and networks are communicating, is uniquely positioned to assess the cause, development, and treatment response of this disorder. Indeed, emerging investigations leveraging connectomics have furthered our knowledge and understanding of epilepsy^4-7.

Epilepsy is a broad class of disorders resulting in recurring seizures (abnormal electrical activity in the brain)⁸. There are many different causes for various forms of epilepsy, including genetics, acute illness, substance withdrawal, acute injury, and electrolyte imbalance⁸.

Although children and older adults are most susceptible to seizures, they can occur in people of any age group or demographic⁸. However, 80% of people living with epilepsy are from low and middle-income countries. With the limited accessibility to proper medical treatment that characterizes this at-risk demographic, it has been proposed that up to 70% of people with epilepsy could live seizure-free¹.

Individual seizure events are complex and heterogeneous, but generally fit in one of three broad categories⁸:

1. Partial seizures begin with excessive synchronized firing in one portion of the cortex⁷. People undergoing a partial seizure are awake and may display irregular movements, have inappropriate autonomic responses, and experience sensory hallucinations such as visual aura⁹.
   
2. Particularly in adult individuals, these partial seizures can transition into generalized seizures and spread across the entire cortex^8,9. When a seizure begins or spreads over the entire cortex, an individual loses consciousness and exhibits tonic-clonic rhythmic movements.
   
3. The third class of seizures is partial-complex seizures, which cause minimal motor symptoms, but leave the individual in an altered, hallucinatory state of consciousness⁸.

Seizures largely result from an imbalance of the major inhibitory-excitatory neurotransmitter systems (GABA and Glutamate respectively) causing a cascade of neural activity and resulting in one of the three seizure types discussed above. However, the neural patterns of epileptic events are widespread, and the specific symptoms which present alongside seizures can only be fully understood by considering which brain networks are compromised^10,11.

As generalized seizures engage the entire cortex, six of the seven major brain networks have been shown to play a role in their manifestation:

• Language: The language network is responsible for processing and generating spoken and written language. TLE patients often present with language impairments and dysfunctional language network connectivity¹². 
• Salience: The salience network helps the brain decide what to focus on. Heightened salience network connectivity is found in TLE patients as well as childhood absence seizures^13,14. 
• Sensorimotor: The sensorimotor network is responsible for motor control and somatic sensations like touch. This network is engaged during some kinds of seizure activity. Moreover, it is also an epileptogenic source in some individuals¹⁵. 
• Central-executive: The central-executive network is the site of higher-order decision-making and external thinking. Altered Central Executive Network connectivity in chronic epilepsy patients is believed to underpin reduced working memory performance and associated cognitive symptoms^16,17. 
• Limbic: The limbic network is responsible for “older” brain functions like memory, anxiety, and fight or flight responses. TLE patients often display impaired functional connectivity within the limbic network, lateralized to the hemisphere nearest to the seizure site¹⁸.
• Default mode: The default mode network (DMN) is engaged when someone is at rest or daydreaming. Loss of consciousness in seizures may be caused by inhibition of systems that keep the DMN active whilst awake¹⁹.

Medication

A wide range of anti-epileptic medications have been developed which counter and suppress the onset of seizures in epilepsy. Broadly, these medications work by suppressing neural activity, primarily within epileptogenic foci and networks.

While these medications are effective in 60-70% of patients in abolishing recurrent seizures, they can consequently induce negative psychological and physiological side effects^22,23. In the 30-40% of drug-resistant epilepsy patients in which these medications are ineffective, two alternate treatment options exist: surgical resection, and vagal nerve stimulation.

Surgery

The most prominent Resection surgery for treatment-resistant epilepsy is Laser Interstitial Thermal Therapy (LITT). This minimally-invasive surgery requires ablating discrete regions of epileptogenic foci, and can give patients life-changing relief from their seizures. However, successful remission of seizures is not guaranteed, and personalizing treatment strategy is quickly becoming a goal of modern neurosurgery.

Solutions such as connectomics have emerged as useful for epilepsy patient-selection and identifying those likely to respond to resective surgery, and will be discussed in the following sections²⁴.

Neuromodulation

Alternative to neurosurgery, deep brain stimulation (DBS) and vagus nerve stimulation (VNS) seeks to treat epilepsy symptoms by regulating neurological activity to the brain. These treatments are generally less invasive than resective and ablative surgery, reducing costs associated with treatment as well as recovery time. However, their effectiveness is highly variable between individuals, and produces a gradual response, reducing seizures to a greater extent over time from beginning treatment^25,26.

Connectomics holds utility in neuromodulatory treatments of epilepsy in both identifying accurate targets for DBS²⁷, as well as tracking the treatment response through changes in underlying brain function over time²⁶.

Figure 2: Common epilepsy DBS treatment sites

With modern neuroimaging techniques such as connectomics, we can examine whole-brain connectivity in the context of epilepsy. Using this technology alongside electroencephalography (EEG) and electrophysiology (EP) makes it possible to pinpoint the brain networks and subsystems affected by or causing epilepsy-related symptoms^11,28. 

Due to the varying nature of the disorder between individuals, personalized brain mapping may be a potential solution for identifying the worst-affected regions. For example, in a 2016 study by Donos et al., the authors paired intracranial electric stimulation with both functional and structural connectivity to investigate epilepsy-related brain changes²⁹. Their combined map consisted of 70 regions per hemisphere outside the epileptic source, which could then be used to assess changes in neural function during epilepsy²⁹. Such applications could help plan intractable epilepsy care. 

Connectomic analyses can also be used to identify potential mechanisms that underlie epileptic subtypes. For example, TLE patients display reduced connectivity between cortical and subcortical networks, while individuals with benign epilepsy and centrotemporal spike often display greater connectivity between cortical networks only³⁰. With the utilization of connectomics, we may generate a better understanding of the connectivity characteristics which define epileptic subtypes, producing greater diagnostic accuracy and further treatments^31-33.

Recent research leveraging connectomic techniques not only demonstrates its potential in identifying where epilepsy is located in the brain, but also in predicting who will respond to different types of treatments^24,34,35.

Alongside seizures, epilepsy has numerous cognitive symptoms, such as impaired motor, emotional, and executive function. These symptoms are directly tied to brain networks, namely the sensorimotor, limbic, and central-executive networks. Indeed, comorbid depression afflicts between 20-55% of epilepsy patients, and has shown to be directly tied with limbic system dysfunction³⁶. 

However within the context of epilepsy, brain networks and symptoms outside of seizures themselves have received relatively little attention. With advanced neuroimaging and connectomics, it may be possible to identify and better understand the brain network alterations associated with non-seizure components of epilepsy, enabling a more tailored and personalized continuum of epilepsy treatment and care^35,36. 

Connectomics enables a detailed structural and functional view of brain changes associated with the wide variety of symptoms which afflict epilepsy patients. Already demonstrating both predictive and therapeutic value, connectomic-approaches are beginning to be used to better understand and treat the symptoms of epilepsy, and in the future may help progress personalized treatment options^{6, 13, 24, 27, 29, 34, 36}.