Human reconsolidation has no single readout

One memory, two answers

A volunteer can stop startling at a learned threat while still expecting the shock.

In a human fear-conditioning experiment, Kindt, Soeter and Vervliet gave participants propranolol before briefly reactivating a conditioned fear memory. Later, the differential fear-potentiated startle response disappeared, yet participants retained explicit knowledge about which cue predicted the shock.

Calling that memory erased compresses two results into one. A physiological expression of fear changed. Reported expectancy did not. The experiment therefore raises a question that follows human reconsolidation research everywhere: which part of the memory is the memory?

The rat result behind the metaphor

The influential biological evidence came from a more direct intervention. In 2000, Nader, Schafe and LeDoux reactivated a conditioned fear memory in rats, then infused the protein-synthesis inhibitor anisomycin into the amygdala. The rats subsequently showed less freezing to the conditioned tone.

The interpretation was that retrieval had made the established memory labile. Restabilizing it required new protein synthesis. Blocking that process after retrieval impaired its later expression.

That experiment earned the term reconsolidation a cellular meaning. It did not establish that every act of human remembering opens an autobiographical memory for revision.

What human experiments can see

Human studies usually cannot observe destabilization itself. They infer it from a pattern: reactivate a memory, introduce a drug or new learning, then test whether some later response has changed. Elsey, Van Ast and Kindt’s 2018 critical review emphasizes that this inference depends on the procedure and the outcome measure.

Even closely related procedures do not always produce the expected result. In 2017, Schroyens, Beckers and Kindt administered propranolol before memory reactivation. They did not find the expected reduction in fear-potentiated startle.

That null result does not, by itself, show that reconsolidation never occurs. It could mean the memory was not destabilized, the intervention did not alter it, or the chosen test did not capture the alteration. Without an independent marker of destabilization, those possibilities are difficult to separate.

The evidence stops at the assay

Here the evidence gives way to interpretation.

The studies support a narrower claim than the popular metaphor. Under some laboratory conditions, retrieving a human memory before an intervention is followed by a durable change in a measured response. They do not show that recall routinely rewrites the whole memory, or that every component changes together.

So the revealing question is not simply whether a recalled memory changed. It is what researchers measured afterward: startle, skin conductance, verbal expectancy, confidence or later recall. One readout can move while another remains.

Until destabilization can be identified independently of its supposed consequences, human reconsolidation remains partly an inference from those consequences. The unresolved problem is not just what recall does to a memory. It is how many different things are being called the memory.