Personalized Closed-Loop Brain Stimulation for Patients With Mild Cognitive Impairment

This research study investigates the feasibility and efficacy of a personalized, closed-loop electroencephalogram-transcranial electrical stimulation (EEG-tES) intervention for individuals with mild cognitive impairment (MCI), addressing the inconsistent results of generic brain stimulation protocols. By integrating artificial intelligence (AI)-derived insights with real-time data, the study aims to customize transcranial electrical stimulation (tES) parameters, including electrode placement, intensity, and frequency to target the specific brain regions responsible for abnormal signaling in each participant. Over the intervention period paired with computerized cognitive training, the project will evaluate improvements in learning, memory, and functional connectivity, while simultaneously identifying clinical and physiological predictors to determine the viability of transitioning this low-cost, non-invasive technology into a remotely supervised, home-based therapy setting. The study duration will be a total of 6-8 weeks.

Study Overview

• Status: Not yet recruiting
• Conditions: MCI
• Intervention / Treatment: Device: Active tES; Behavioral: Cognitive Interventions

• Study Type: Interventional
• Enrollment (Estimated): 10
• Phase: Not Applicable

Contacts and Locations

Study Locations

• United States
  • Maryland
    • Baltimore, Maryland, United States, 21287
      • Johns Hopkins Hospital
      • Contact: Kyrana Tsapkini, PhD; Phone Number: 410-736-2940; Email: tsapkini@jhmi.edu

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: Yes

Description

Inclusion Criteria:

• Must be between 45-85 years of age.
• Must be proficient in English.
• Must have a minimum of high-school education.
• Must be diagnosed with 'probable Alzheimer's Disease (AD)' in specialized diagnostic centers with neuropsychological (e.g., RAVLT) and AD biomarkers according to 2011 guidelines.
• The investigators will also perform extensive testing in the investigators' test battery including the Mnemonic Similarity Test (MST) that discriminates and measures the most salient hippocampal deficit-pattern separation (PS).

Exclusion Criteria:

• People with previous neurological disease including vascular dementia (e.g., stroke, developmental dyslexia, dysgraphia or attentional deficit).
• People with hearing loss (> 25 decibel, using audiometric hearing screen).
• People with uncorrected visual acuity loss.
• People with advanced dementia or severe language impairments (MMSE <15, or Montreal Cognitive Assessment <10, or language Frontotemporal Dementia-specific Clinical Dementia Rating (FTD-CDR) =3).
• People with pre-existing psychiatric disorders such as behavioral disturbances, severe depression, or schizophrenia that do not allow these people to comply or follow the study schedule and requirements such as repeated evaluation and therapy.

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Treatment
• Allocation: N/A
• Interventional Model: Single Group Assignment
• Masking: None (Open Label)

Number of Arms

1

Arms and Interventions

Participant Group / Arm

Intervention / Treatment

Experimental: Personalized closed-loop tES treatment + Cognitive Interventions; Participants will receive personalized, closed-loop active tES for 3 weeks. Active tES will be administered for 20 minutes. Cognitive training exercises will be administered concurrently with the tES and will take approximately 30 minutes to complete.

Device: Active tES; This three-week, closed-loop program integrates personalized tES with computerized cognitive training to drive neuroplasticity. Following baseline electroencephalogram (EEG) mapping, participants undergo daily 30-minute sessions consisting of four "loops." Each loop begins with an AI-analyzed EEG to calibrate stimulation parameters, followed by 5-minute stimulation periods.; Behavioral: Cognitive Interventions; In between tES treatment loops, the participants will be asked to complete cognitive exercises. By pairing real-time brain modulation with targeted executive function exercises, the intervention aims to improve language skills and clinical outcomes through precise, data-driven cortical targeting.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Change in auditory recall accuracy based on the sum of words recalled in Trials 1-5 of semantically related - trained word-lists	Each trained word-list (practiced during the intervention period) will consist of 12 semantically related words (e.g., birds). Word lists will be constructed using psycholinguistic databases. There will be 5 Trials to learn each list. The investigators will compute the raw score of items correctly recalled by summing all scores from Trial 1 to Trial 5 and transforming to percent correct (range: 0-100%) at each time point of the study. Higher score is better. Increase in scores is considered a benefit.	Before intervention, immediately after intervention and 1 month post intervention
Change in auditory delayed recall accuracy of semantically related - trained word-lists	Each trained word-list (practiced during the intervention period) will consist of 12 semantically related words (e.g., birds). Word lists will be constructed using psycholinguistic databases. There will be 5 Trials to recall each list, and then participants will be asked to recall that list 20 minutes later (delayed recall). The investigators will compute the raw score of items correctly recalled (delayed recall) and transform to percent correct (range: 0-100%) at each time point of the study. Increase in scores is considered a benefit.Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in auditory recall accuracy based on the sum of words recalled in Trials 1-5 of semantically unrelated - trained word-lists	Each trained word-list (practiced during the intervention period) will consist of 12 semantically unrelated words as in Rey Auditory-Verbal Learning Test (RAVLT). Word lists will be constructed using psycholinguistic databases. There will be 5 Trials to learn each list. The investigators will compute the raw score of items correctly recalled by summing all scores from Trial 1 to Trial 5 and transforming to percent correct (range: 0-100%) at each time point of the study. Increase in scores is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in auditory delayed recall accuracy of semantically unrelated - trained word-lists	Each trained word-list (practiced during the intervention period) will consist of 12 semantically unrelated words (as in RAVLT). Word lists will be constructed using psycholinguistic databases. There will be 5 Trials to recall each list, and then participants will be asked to recall that list 20 minutes later (delayed recall). The investigators will compute the raw score of items correctly recalled (delayed recall) and transform to percent correct (range: 0-100%) at each time point of the study. Increase in scores is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in auditory recall accuracy based the sum of words recalled in Trials 1-5 of semantically related - untrained word-lists	Each untrained word-list (not practiced during the intervention period) will consist of 12 semantically related words (e.g., birds). Word lists will be constructed using psycholinguistic databases. There will be 5 Trials to learn each list. The investigators will compute the raw score of items correctly recalled by summing all scores from Trial 1 to Trial 5 and transforming to percent correct (range: 0-100%) at each time point of the study. Increase in scores is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in auditory delayed recall accuracy of semantically related - untrained word-lists	Each untrained word-list (not practiced during the intervention period) will consist of 12 semantically related words (e.g., birds). Word lists will be constructed using psycholinguistic databases. There will be 5 Trials to recall each list, and then participants will be asked to recall that list 20 minutes later (delayed recall). The investigators will compute the raw score of items correctly recalled (delayed recall) and transform to percent correct (range: 0-100%) at each time point of the study. Increase in scores is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in auditory recall accuracy based on the sum of words recalled in Trials 1-5 of semantically unrelated - untrained word-lists	Each untrained word-list (not practiced during the intervention period) will consist of 12 semantically unrelated words (as in RAVLT). Word lists will be constructed using psycholinguistic databases. There will be 5 Trials to learn each list. The investigators will compute the raw score of items correctly recalled by summing all scores from Trial 1 to Trial 5 and transforming to percent correct (range: 0-100%) at each time point of the study. Increase in scores is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in auditory delayed recall accuracy of semantically unrelated - untrained word-lists	Each untrained word-list (not practiced during the intervention period) will consist of 12 semantically unrelated words (as in RAVLT). Word lists will be constructed using psycholinguistic databases. There will be 5 Trials to recall each list, and then participants will be asked to recall that list 20 minutes later (delayed recall). The investigators will compute the raw score of items correctly recalled (delayed recall) and transform to percent correct (range: 0-100%) at each time point of the study. Increase in scores is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Change in Rey Auditory-Verbal Learning Test (RAVLT) score	RAVLT is a well-established verbal memory test. RAVLT includes a 5-trial presentation of a 15-word list (List A), a single presentation of an interference list (List B)(Trial 6), two post-interference recall trials (one immediate - Trial 7, one delayed - Trial 8) and recognition of the target words in the orthographic modality with distractors (Trial 9). Scoring includes the percent score of Trial 1, Trial 5, Trial 8 and Trial 9 as well as the sum of Trial 1 through 5, and the difference between Trial 5 and Trial 1 computed as the percent difference between the scores before intervention and each time point after. Increase in score is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in Mini Mental State Examination (MMSE)	MMSE is a well-established cognitive assessment test. It examines functions including registration (repeating named prompts), attention and calculation, recall, language, ability to follow simple commands and orientation. The total raw score is out of 30 points. The investigators will compute the raw score of items correct and transform to percent correct (range: 0-100%), computing change in outcome in percent difference between before intervention and each time point after. Increase in score is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in Mnemonic Similarity Task (MST) score	MST is a well-established test in order to assess high interference memory and general recognition memory via pattern separation. It involves differentiating between previously learned images and novel images. For the MST tasks, the Pattern Separation (PS) score will be calculated using two measures: a) the rate of similar items correctly identified minus the rate of similar items misidentified as new (S|S-S|N); b) the rate of similar items correctly identified minus the rate of similar items misidentified as old (S|S-O|S). The number of correct responses for each category of items (i.e., old, similar, new) and the type of errors (i.e., identifications of new items as similar; identification of similar items as old) will also be tracked. Change in outcome in percent difference will be computed between the scores before intervention and each time point after. Increase in scores is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in word repetition score	Temple Assessment of Language and Short-Term Memory in Aphasia (TALSA) tasks include word repetition, with sets of 1-6 words. Scoring will be based on percent of words correctly repeated. The investigators will compute the raw score of items correct and transform to percent correct (range: 0-100%), computing change in outcome in percent difference between before intervention and each time point after. Increase in score is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in non-word repetition score	TALSA tasks include non-word repetition, with sets of 1-6 non-words. Scoring will be based on percent of non-words correctly repeated. The investigators will compute the raw score of items correct and transform to percent correct (range: 0-100%), computing change in outcome in percent difference between before intervention and each time point after. Increase in score is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in sentence repetition score	Sentence repetition tasks come from the TALSA, with scoring based on percent of words in sentences correctly repeated. The investigators will compute the raw score of items correct and transform to percent correct (range: 0-100%), computing change in outcome in percent difference between before intervention and each time point after. Increase in score is considered a benefit.Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in oral naming Boston Naming Test score	Accuracy in oral picture naming (30-item Boston Naming Test). The Boston Naming Test is a widely used picture naming test that detects lexical retrieval deficits in the oral modality. The investigators will compute the raw score of items correct and transform to percent correct (range: 0-100%), computing change in outcome in percent difference between before intervention and each time point after. Increase in score is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in oral naming Philadelphia Naming Test score	Accuracy in oral picture naming (Philadelphia Naming Test). The Philadelphia Naming Test is an extensive picture naming test that comprises 275 items from a wide range of frequencies and other psycholinguistic characteristics. The investigators will compute the raw score of items correct and transform to percent correct (range: 0-100%), computing change in outcome in percent difference between before intervention and each time point after. Increase in score is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in written naming as assessed by Boston Naming Test	Accuracy in written picture naming (30-item Boston Naming Test). The investigators will compute the raw score of items correct and transform to percent correct (range: 0-100%), computing change in outcome in percent difference between before intervention and each time point after. Increase in score is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in written naming as assessed by Philadelphia Naming Test	Accuracy in written picture naming (Philadelphia Naming Test). The investigators will compute the raw score of items correct and transform to percent correct (range: 0-100%), computing change in outcome in percent difference between before intervention and each time point after. Increase in score is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in oral naming of action as assessed by Hopkins Assessment of Naming Actions (HANA)	Accuracy in oral naming of actions. The investigators will compute the raw score of items correct and transform to percent correct (range: 0-100%), computing change in outcome in percent difference between before intervention and each time point after. Increase in score is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in syntactic comprehension as assessed by Subject-relative, Object-relative, Active, Passive (S.O.A.P.) Syntactic Battery	The 40-item Subject-relative, Object-relative, Active, Passive (S.O.A.P.) Syntactic Battery of various sub-tests will be used to assess argument structure comprehension and production. The investigators will compute the raw score of items correct and transform to percent correct (range: 0-100%), computing change in outcome in percent difference between baseline and each time point. Increase in score is considered benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in verbal fluency task score	Verbal fluency tasks (semantic and letter fluency) involve generating as many words as possible in one minute. Scoring will be based on number of words generated per minute. The investigators will compute the raw score of items correct and compute change in outcome between baseline and each time point. Increase in score is considered benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in spelling as assessed by the Johns Hopkins Dysgraphia battery	Accuracy in spelling using the Johns Hopkins Dysgraphia battery. The investigators will compute the raw score of items correct using a spelling scoring system accounting for additions, substitutions, and deletions, and transform to percent correct (range: 0-100%), computing change in outcome in percent difference before intervention and each time point after. Increase in score is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in digit span forward score	Digit span forward involves the recall of a series of single digits (sets of 1-8 digits) in the same order the digits were presented. Scoring will be based on the number of consecutive digits correctly recalled. The investigators will compute the change in outcome between the time point before intervention and each time point after. Increase in score is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in digit span backward score	Digit span backward involves the recall of a series of single digits (sets of 1-8 digits) in the reverse order than the digits were presented. Scoring will be based on the number of consecutive digits correctly recalled. The investigators will compute the change in outcome between the time point before intervention and each time point after. Increase in score is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in spatial span forward score	Spatial span forward involves the recall of a series of positions on a board (sets of 1-9) in the same order the digits were presented. Scoring will be based on the number of consecutive positions correctly recalled. The investigators will compute the change in outcome between the time point before intervention and each time point after. Increase in score is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in spatial span backward score	Spatial span backward involves the recall of a series of positions (sets of 1-8) in the reverse order than the digits were presented. Scoring will be based on the number of consecutive positions correctly recalled. The investigators will compute the change in outcome between the time point before intervention and each time point after. Increase in score is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in semantic content of connected speech (Cookie Theft)	Using the Cookie Theft image from the Boston Diagnostic Aphasia Examination (BDAE) investigators will obtain representative language samples as participants describe the image. The investigators will combine the raw score of all items (semantics) correct from the image and transform to percent correct (range: 0-100%), computing change in outcome in percent difference between before intervention and each time point after. Increase in score is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in semantic content of connected speech (Circus)	Using the Circus image from the Apraxia Battery for Adults (ABA) investigators will obtain representative language samples as participants describe the images. The investigators will combine the raw score of all items (semantics) correct from the image and transform to percent correct (range: 0-100%), computing change in outcome in percent difference between before intervention and each time point after. Increase in score is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention
Change in attention and manipulation of information scores	Using the Trail Making Test (TMT) parts A and B, which include the sequential connection of letters/numbers in order to complete a trail, the investigators will obtain the time required by the participants to finish the tasks. Decrease in the time is considered a benefit. Higher score is better.	Before intervention, immediately after intervention and 1 month post intervention

Collaborators and Investigators

• Sponsor: Johns Hopkins University
• Collaborators: Mackler-Goding Foundation
• Investigators: Principal Investigator: Kyrana Tsapkini, PhD, Johns Hopkins University

Study record dates

Study Major Dates

• Study Start (Estimated): June 1, 2026
• Primary Completion (Estimated): August 1, 2026
• Study Completion (Estimated): August 1, 2026

Study Registration Dates

• First Submitted: March 18, 2026
• First Submitted That Met QC Criteria: April 1, 2026
• First Posted (Actual): April 6, 2026

Study Record Updates

• Last Update Posted (Actual): May 8, 2026
• Last Update Submitted That Met QC Criteria: May 6, 2026
• Last Verified: May 1, 2026

More Information

Terms related to this study

• Keywords: MCI; EEG; Closed-loop; Personalized treatment; tES
• Other Study ID Numbers: IRB00553664

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: Yes