What Happens Between Learning, Forgetting, and Learning Again?

Is the second encounter with the same information still the same encounter? What if retrieving a memory also depends on the states an organism can re-enter?

Associative Memory · EEG Microstates · Neural Reinstatement · Body-Territory · Connectomic Reserve · Metanoia

Imagine recognizing someone and remembering where you met, but being unable to recall their name. A few hours later, another person mentions that name, and the memory immediately returns.

What changed? Was the information forgotten, or were the conditions necessary for retrieving it temporarily unavailable?

Now imagine that during your next encounter, you remember the name spontaneously.

Is the second encounter with the same information still the same encounter?

This question finds a valuable scientific opportunity in the work of Gi-Hwan Shin, Young-Seok Kweon, Ha-Na Jo, Heon-Gyu Kwak, and Seong-Whan Lee, from Korea University.

In October 2026, the researchers published An EEG dataset for multi-state memory processing in an associative word-pair task in Scientific Data.

Their contribution was to make recordings available for investigating how brain activity varies across encoding, retrieval, re-encoding, and resting states.

1. Learning is more than storing information

Memory does not function like a file that remains equally accessible under every condition.

During encoding, the organism establishes associations related to an experience. During retrieval, internal and external cues may facilitate the reactivation of those representations.

However, an incorrect response does not necessarily mean that the information has disappeared.

A memory may become less accessible because retrieval cues are inadequate, interference is present, or the conditions during retrieval differ from those experienced during learning.

Neuroscience investigates these possibilities through concepts such as state-dependent memory and neural reinstatement.

This changes our question: beyond asking whether someone remembers, we need to investigate under which conditions remembering becomes possible.

2. How did Shin and colleagues organize the experiment?

The dataset includes 41 participants performing an associative memory task involving word pairs.

The protocol included questionnaires, resting periods, and encoding, retrieval, and re-encoding phases.

During encoding, participants encountered associations between words.

During retrieval, they attempted to recall previously presented information.

During re-encoding, they had another opportunity to process the same associations.

The researchers provided recordings from 28 EEG channels and four electrooculography channels, together with behavioral responses, reaction times, and subjective difficulty ratings.

This organization allows different cognitive states to be compared within the same experimental protocol.

However, the publication is a Data Descriptor: the authors documented and shared data rather than establishing causal mechanisms of memory consolidation.

3. Can averages hide the moment when remembering happens?

In Blogs 03 and 06, we discussed how average measurements can conceal important temporal characteristics.

The investigation by Rayson and colleagues on beta bursts identified associations between specific features of these events and children's motor abilities.

We can formulate a similar question about memory.

Could two participants show comparable average EEG power but different temporal patterns during retrieval?

Perhaps one presents brief changes around the retrieval cue, while another displays a more distributed organization of activity.

This does not mean that a particular oscillation contains a memory.

The challenge is to discover when, where, and under which conditions brain activity participates in retrieval.

4. EEG Microstates: the brain as a sequence of configurations

One particularly interesting analytical approach involves EEG Microstates.

These are relatively stable topographic configurations of electrical activity across the scalp, typically lasting approximately 50–120 milliseconds before transitioning into other patterns.

The 2026 review EEG microstates: from methodological foundations to clinical translation discusses their applications and limitations.

Researchers can investigate duration, occurrence, temporal coverage, and transition probabilities.

BrainLatam has already explored this perspective in EEG Microstates — The Brain as a Sequence of Small Transitions.

However, microstates are not frequency bands, nor do they automatically represent specific thoughts or memories.

Their value lies in describing how different configurations succeed one another during cognitive processing.

5. What new questions can Shin's dataset help investigate?

We could examine whether microstate duration and coverage differ between encoding, retrieval, and re-encoding.

It may also be possible to compare sequences preceding correct and incorrect responses.

Another question concerns similarity: when an association is retrieved, do configurations observed during its original encoding reappear?

The 2025 study Study-phase reinstatement predicts subsequent recall identified relationships between reinstatement of neural patterns and subsequent memory using intracranial EEG.

That finding provides a scientific rationale for investigating reinstatement in Shin's recordings.

However, identifying similar patterns does not independently demonstrate the reactivation of an identical cellular engram.

Analyses would need to control for artifacts, individual differences, and the characteristics of each experimental phase.

6. Can the same memory be accessed through different states?

A second possibility emerges here.

Successful retrieval may not require reproducing exactly the brain configuration present during learning.

Through additional experiences, an organism may develop different cues and strategies capable of facilitating access to the same information.

Research on expert memory, such as Practicing perfection: Piano performance as expert memory by Roger Chaffin and Gabriela Imreh, demonstrates the importance of structured retrieval cues in highly practiced skills.

This opens two distinct hypotheses.

The first predicts that retrieval is facilitated by reinstating previous configurations.

The second predicts greater flexibility in retrieving information through different functional configurations.

Expertise could strengthen the first possibility, expand the second, or combine both.

These alternatives must be investigated rather than assumed.

7. From the brain to the Body-Territory

The decolonial neuroscience perspective proposed by BrainLatam seeks to expand the unit of investigation: from the isolated brain toward the Body-Territory.

The organism attempting to remember also breathes, regulates circulation, experiences emotional states, and carries a history of relationships and previous encounters.

Language may provide contextual retrieval cues, but it represents only one possible dimension.

We could investigate whether EEG patterns, respiratory phase, autonomic regulation, and environmental conditions are associated with memory accessibility.

This does not mean that all these processes must be synchronized.

The question is which similarities or differences between encoding and retrieval states influence performance.

Shin's dataset does not contain the respiratory and autonomic measurements required to test this complete hypothesis.

8. Glia and Ca²⁺: what EEG cannot observe

Electrical activity recorded at the scalp provides only one window into the processes involved in memory.

In Astroengrams: rethinking the cellular substrate for memory, Javier Sánchez Romero and Marta Navarrete discuss evidence that astrocytic ensembles contribute to processes associated with memory formation and retrieval.

This broadens the investigation toward cellular mechanisms involving Ca²⁺ signaling, synaptic regulation, and interactions between neurons and glial cells.

However, EEG does not directly measure intracellular calcium or identify individual astrocytes.

Complementary cellular investigations would be necessary to study these mechanisms.

The Body-Territory perspective does not eliminate experimental boundaries; it requires recognizing what each method observes and what remains beyond its reach.

9. Connectomic Reserve and Metanoia of the Metastable Brain

The hypothesis proposed by Roberto Lent, Fernanda Tovar-Moll, and Diego Szczupak suggests that relatively underused anatomical connections may contribute to neuroplasticity.

This inspires the question of whether learning expands the functional possibilities available for retrieving particular information.

However, greater microstate diversity does not demonstrate recruitment of previously silent anatomical connections.

In BrainLatam's proposal of Metanoia of the Metastable Brain, the focus is on persistent changes in an organism's repertoires and possibilities for transitioning between states.

An A→B→A design could compare initial processing, an intervening experience, and the subsequent encounter.

Delayed retention assessments would be necessary to distinguish temporary changes from lasting learning.

Metastability would also require specific dynamical measurements; it cannot be established merely by observing transitions between microstates.

10. What experiment could we propose next?

The first step would be to reanalyze Shin's EEG recordings, comparing microstates, pattern similarity, behavioral responses, and reaction times.

A second step would involve new recordings combining EEG, respiration, ECG, and subjective assessments.

Researchers could manipulate retrieval conditions while keeping the learned information constant.

The objective would be to determine whether particular neural and physiological configurations facilitate remembering and whether successive experiences expand the conditions under which a memory remains accessible.

Shin and colleagues' original contribution remains central: an open dataset that enables researchers to investigate different states of memory processing within the same protocol.

BrainLatam adds a question concerning the organism as a whole:

Does learning again merely strengthen a memory, or does it also expand the Body-Territory states from which that memory can be retrieved?

The second encounter may not be the same because the organism already carries the history of the first.

In Blog 08, we will investigate priming: how previous experiences prepare our expectations even before the next stimulus appears.

Meta description: A 2026 EEG dataset of 41 participants opens new questions about memory retrieval, microstates, neural reinstatement, relearning, and the Body-Territory.

Scientific note: Shin and colleagues' contribution is an open EEG dataset, not a causal demonstration of memory consolidation or retrieval mechanisms. The proposed microstate analyses, relationships with physiological states, and connections to Connectomic Reserve and Metanoia of the Metastable Brain remain research hypotheses.

BRAINLATAM | W47/2026 | BLOG 07/12 — ENGLISH VERSION