In this study, participants will receive unilateral Deep Brain Stimulation (DBS) for treatment of epilepsy, with network-based stimulation targets specifically defined using a stereo-electro-encephalographic evaluation and chronic recordings using the Medtronic Percept™ primary cell (PC) Neurostimulator DBS System with BrainSense™ Technology. The hypothesis is that, compared to no stimulation or to standard duty cycle high frequency stimulation, epilepsy neuromodulation using low frequency stimulation and informed by network architecture in patients with epilepsy that arises in a hippocampus that also subserves memory - epilepsy in a precious hippocampus (EPH) - will result in a significant decrease in seizure frequency and severity, paralleled by a decrease in EEG spike counts and improved memory function.

Different stimulation types will be administered in a crossover fashion, as follows. There will be four four-month periods of low-frequency DBS stimulation, and in each of these four-month periods, stimulation will occur at one of four different sites [the anterior nucleus of the thalamus (ANT), entorhinal cortex (ERC), piriform cortex (PiC), and hippocampal fornix (HCF)], with the order of receipt differing among study participants. There will be a 3-month washout period after each 4-month stimulation period, with the washout being standard of care (SOC) high-frequency DBS stimulation of the ANT. Finally, there will be a 7 to 12 month DBS stimulation period with the stimulation type that yielded the best results.

Inclusion Criteria:

• Patients with a presumptive diagnosis of EPH determined by the group of clinicians who participate in patient management conference.
• Ability to comply with test directions and provide informed consent or assent to the study, i.e. cognitively able to participate in studies [typically intelligence quotient (IQ) of 65 or above].
• Relatively preserved verbal memory - as determined via formal neuropsychological evaluation performed by the neuropsychologist. The values must within 1.5 standard deviation (SD) of the mean for verbal memory
• Proficient in English, as all of our tasks and consent forms will be in English and the inclusion of non-English speakers will introduce another confound in this small sample size and preclude grouped analysis
• Age 18 - 65 years (we expect the trial to take 5 years and wish to target patients with minimal medical co-morbidities)
• Must have a minimum of 2 seizures of any type per month - this is essential to be able to detect the impact of neuromodulation on the epilepsy over relatively short intervals of time. Patients with secondary generalized seizures may also be enrolled so long as they have a maximum of 20 generalized seizures in the past 12 months (prior to enrollment), or an average of no more than 3 generalized seizures per month.

Exclusion Criteria:

• Impaired reading and cognitive functions (more than 3 standard deviations below the mean, usually an IQ < 60), as determined by preoperative neuropsychological testing.
• Patients with gross structural abnormalities (hamartomata, tumors, vascular malformations, diffuse malformations of cortical development) in the brain that raise the possibility of dual pathology resulting in the epilepsy and by derivation, a larger epilepsy network.
• Patients with neurological conditions such as recent history (within past 5 years) of a stroke, encephalitis and meningitis. Any patient with a current diagnosis of these conditions will also be excluded.
• Patients with any episodes of status epilepticus in the past 12 months prior to enrollment.
• Patients with uncontrolled prominent psychiatric comorbidity that will preclude their meaningful participation.
• Patients with a Beck Depression Inventory II score at baseline examination greater than or equal to 29 (i.e., severe depression).
• Patients who have attempted suicide in the past 12 months.
• Patients with memory impairment due to other neurological conditions such as dementia and Parkinson's disease.
• Patients who are unable to speak or comprehend English. The inclusion of multiple languages will make task development and grouped comparisons of neuro-psychology data difficult.
• Patients with cardiac pacemakers, intracranial aneurysm clips, or other potentially mobile implanted metallic devices that are deemed MRI incompatible by the manufactures. The absence of high resolution structural imaging precludes appropriate targeting of the regions of interest.
• Profound hippocampal sclerosis with prominent atrophy of the majority of the hippocampus (equivalent to ILAE type III).
• Prior brain surgery for any reason or failed prior brain neuromodulation [prior vagus nerve stimulation (VNS) therapy is acceptable so long as it is held constant for the duration of the trial].
• History of or current non-epileptic spells (will confound accuracy of seizure detection with ANT Percept PC and the precision of the estimate of the neuromodulation effect).
• Patients who are pregnant. All female participants of childbearing potential will be counselled prior to enrollment regarding the unknown risks of treatment on a fetus and the importance of using contraception while they are a subject in this study. If a female participant becomes pregnant during the study, they will returned to FDA-approved ANT stimulation parameters (standard of care).