A meta-analysis suggests that TMS targeting the hippocampal network selectively improves episodic memory

This meta-analysis demonstrates that Hippocampal Indirectly Targeted Stimulation (HITS) robustly and selectively enhances episodic memory, particularly recollection-based performance, in both healthy and clinical populations by targeting the hippocampal network without serious adverse events.

The Gist

Imagine your brain is a bustling city. Deep inside this city lies a special, ancient library called the Hippocampus. This library is responsible for writing down new stories about your life—what you had for breakfast, the name of a new neighbor, or the plot of a movie you saw yesterday. This is what scientists call episodic memory.

Sometimes, the roads leading to this library get clogged, or the library itself gets a bit rusty. This happens in conditions like aging or Alzheimer's disease, making it hard to remember new things.

For a long time, scientists wanted to fix these roads without doing surgery. They found a tool called TMS (Transcranial Magnetic Stimulation). Think of TMS as a "magnetic remote control" that can send gentle pulses to the surface of the brain to wake it up.

The Problem: You Can't Reach the Library

The problem is that the Hippocampus library is buried deep underground. You can't stick the magnetic remote control deep inside the skull without causing damage. It's like trying to fix the engine of a car by tapping on the hood; you can't reach the engine directly.

The Solution: The "Indirect" Shortcut

This paper introduces a clever workaround called HITS (Hippocampal Indirectly Targeted Stimulation).

Instead of trying to zap the library directly, the scientists found a "relay station" on the surface of the brain (specifically in the back-left area, near the parietal lobe) that is directly connected to the library. It's like finding a phone line that connects your house directly to the library's front desk.

By using the TMS remote control on this surface "relay station," they can send a signal that travels down the line and wakes up the deep library.

What Did They Find? (The Big Meta-Analysis)

The authors didn't just do one experiment; they gathered data from 38 different studies involving over 1,000 people. They acted like detectives, piecing together all the clues to see if this "indirect" method actually works.

Here is what they discovered, translated into plain English:

1. It Works Like a Magic Key
When they used this method, people's ability to remember new stories (episodic memory) got significantly better. It was like giving the library a fresh coat of paint and a new set of lights. The improvement was real and consistent across many different groups of people, from young healthy adults to older adults with memory issues.

2. It's a Laser, Not a Flashlight
This is the most important part. The treatment didn't just make people "smarter" in general. It didn't help them pay attention better, solve math problems faster, or learn new languages. It only helped with the specific type of memory that the library handles.

• Analogy: Imagine a flashlight that only lights up the "History" section of a library but leaves the "Science" and "Cooking" sections in the dark. That's how specific this treatment is.

3. The "Recollection" vs. "Recognition" Difference
The study found that the treatment worked best for recollection.

• Recognition is like seeing a face in a crowd and thinking, "I know that person."
• Recollection is like looking at that same face and remembering, "Oh, that's Bob, he lives next door and he loves gardening."
  The treatment was much better at helping people do the deep "Bob" work (recollection) than just the surface "I know him" work. This proves the treatment is hitting the right deep-brain circuits.

4. Timing is Everything
The study found a crucial rule about when to use the remote control:

• Before the task: If you zap the brain before you try to learn something new, it works great. It's like tuning a radio station before you start listening to the news.
• After the task: If you zap the brain after you've already learned something but before you are tested on it, it doesn't really help.
• Takeaway: This suggests the treatment helps your brain encode (write down) new memories, rather than helping you retrieve (find) old ones.

5. It's Safe
Out of over 1,200 people treated, there were no serious side effects. It was as safe as a standard medical checkup.

The "Sweet Spot" for Future Treatments

The researchers also figured out how to make this treatment even better for the future. They found that the best results happen when:

1. You use the treatment before the person tries to learn.
2. You test the person using recollection tasks (asking them to recall details) rather than just recognition.

The Bottom Line

This paper is a huge step forward. It proves that we can use a non-invasive "remote control" on the surface of the brain to fix deep memory problems. It's not a magic cure-all, but it is a highly specific, safe, and effective tool that could one day help people with memory loss regain their ability to remember the stories of their lives.

Think of it as finally finding the right key to unlock the deep library, without ever having to break down the walls.

Technical Summary

1. Problem Statement

Episodic memory, which relies critically on the hippocampus and its associated cortical network, is frequently impaired in neurological and psychiatric disorders (e.g., Alzheimer's disease, mild cognitive impairment). While Transcranial Magnetic Stimulation (TMS) has shown promise in modulating brain networks, the efficacy of Hippocampal Indirectly Targeted Stimulation (HITS)—a method where TMS is applied to neocortical regions (typically parieto-occipital) with high functional connectivity to the hippocampus to indirectly modulate the hippocampal network—has been inconsistent across individual studies.

There was a lack of systematic, quantitative evidence regarding:

• Whether HITS reliably enhances episodic memory across diverse study designs and populations.
• Whether these effects are specific to episodic memory or generalize to other cognitive domains.
• Which specific experimental parameters (e.g., timing, task format, targeting method) optimize therapeutic outcomes.

2. Methodology

The authors conducted a systematic review and meta-analysis of studies published between September 2014 and April 2025.

• Data Collection:

  • Search Strategy: PubMed and Scopus databases were searched using terms "memory" AND "transcranial magnetic stimulation."
  • Inclusion Criteria: Studies using rTMS on hippocampal network locations (parietal/occipital neocortex) with at least one objectively scored episodic memory outcome.
  • Sample: 47 studies met inclusion criteria; 38 provided sufficient data for meta-analysis ($N=1,009$ subjects). These studies yielded 253 statistical comparisons (140 episodic memory effects, 113 non-memory effects).
  • Excluded Studies: 10 studies were excluded due to insufficient statistical data, overlapping samples, or lack of direct performance outcomes.

• Statistical Analysis:

  • Effect Size: Data were transformed into Hedges' g to normalize effect sizes and correct for small sample bias.
  • Modeling: A mixed-effects univariate meta-regression model with random effects by study was used to estimate overall effects.
  • Moderator Analysis: A bootstrapped Lasso regression (1,000 iterations) with cross-validation was used to select significant study design factors from 11 candidates (e.g., Task format, Timing, Targeting, Population, Intensity). Retained factors were then tested in a multivariate meta-regression.
  • Sensitivity: Outlier removal (studentized deleted residuals) was performed to ensure robustness.
  • Bias Assessment: Funnel plots and Egger's regression tests were used to assess publication bias.

3. Key Contributions

• First Comprehensive Meta-Analysis: This is the first quantitative synthesis of HITS effects on episodic memory, covering both healthy and clinical populations.
• Mechanistic Specificity: The study provides robust evidence that HITS effects are selective to episodic memory and, more specifically, to recollection processes, rather than general cognitive enhancement.
• Optimization Guidelines: The analysis identifies specific design factors that significantly modulate efficacy, offering a roadmap for future clinical trials and experimental designs.
• Safety Profile: The review confirms a low incidence of adverse events across a large pooled sample ($N=1,223$).

4. Key Results

A. Overall Efficacy

• Episodic Memory: HITS produced a robust, positive effect on episodic memory performance ($g = 0.44$, $p < 0.001$).
• Non-Memory Tasks: HITS had no significant effect on non-memory cognitive domains (attention, working memory, executive function, language) ($g = 0.04$, $p = 0.12$).
• Selectivity: The difference between memory and non-memory effects was statistically significant ($p < 0.001$), confirming that HITS does not produce general cognitive arousal or off-target benefits.

B. Significant Moderators (Factors Influencing Efficacy)

Three factors were identified as significantly modulating the effect size:

1. Task Format (Recollection vs. Recognition):

   • Effects were significantly larger for tasks measuring recollection (e.g., free recall, cued recall) compared to recognition tasks ($g$ modification = +0.23, $p = 0.004$).
   • This aligns with the neurobiological understanding that the hippocampal network is more critical for recollection than familiarity-based recognition.

2. Timing of Stimulation:

   • Pre-task stimulation (delivered before memory encoding) yielded significantly larger effects ($g$ modification = +0.68, $p < 0.001$).
   • Post-encoding stimulation (delivered immediately after encoding but before retrieval) showed near-zero effects and was associated with the majority of negative (impairing) outcomes.
   • Implication: HITS likely facilitates memory formation (encoding) rather than retention or retrieval.

3. Targeting Method:

   • Individualized targeting (subject-specific MRI connectivity) showed a trend toward better efficacy than group/atlas-based targeting, though this did not survive sensitivity analysis with outlier removal ($p = 0.09$).

C. Optimized Design Projection

Studies incorporating both recollection-format tasks and pre-task stimulation (the "optimized" design) showed a projected effect size of $g = 0.66$, compared to $g = 0.36$ for studies with non-optimized designs.

D. Safety and Populations

• Safety: No serious adverse events were reported across the entire sample ($N=1,223$), including patients with mild-to-moderate Alzheimer's dementia and psychiatric conditions.
• Populations: Positive effects were observed across healthy young adults, older adults, and clinical populations, suggesting broad applicability.

5. Significance and Implications

• Clinical Translation: The findings support HITS as a viable, non-invasive therapeutic strategy for memory disorders, particularly those characterized by recollection deficits (e.g., early Alzheimer's).
• Mechanistic Insight: The selectivity for recollection and the timing dependence (pre-encoding efficacy) provide strong evidence that HITS modulates the specific neural mechanisms of memory formation within the hippocampal network, rather than inducing non-specific arousal.
• Future Research Directions:
  • Future clinical trials should prioritize pre-task stimulation protocols and recollection-based outcome measures.
  • Research should explore whether these effects are maintained over long-term periods (months/years).
  • Further investigation is needed to determine if HITS is effective for disorders involving distinct network pathologies beyond the standard hippocampal network.

In conclusion, this meta-analysis establishes HITS as a robust, selective, and safe method for enhancing episodic memory, with clear guidelines for optimizing experimental and clinical protocols to maximize therapeutic benefit.