BGPT: Paper Review: Long-term memory underlying hippocampus-dependent social recognition in mice

Explore by Goal

Quick Explanation Copied

Core claim

Group-housed adult male mice show long-term social recognition (≥1–7 days) that depends on hippocampal integrity, de novo protein synthesis, and CREB α/δ function; social isolation disrupts long-term—but not 30-min—memory.

Long Explanation

Paper Review (visual first)

Title: Long-Term Memory Underlying Hippocampus-Dependent Social Recognition in Mice (accepted Sep 1, 1999; published 2000) DOI: 10.1002/(SICI)1098-1063(2000)10:1<47::AID-HIPO5>3.0.CO;2-6

What the experiment measures (behavioral readout)

Social recognition index: adults investigate juveniles; investigation time decreases when the juvenile is familiar at test.
Recognition ratio: test investigation duration divided by (initial + test); 0.50 ≈ no recognition, while <0.50 indicates recognition.
Designed as a within-subject “initial vs test” structure, with group comparisons using ANOVA/paired t-tests; collection was reportedly blind to genotype/treatment.

Recognition ratio across delays (familiar vs novel)

Values come directly from the paper’s recognition-ratio table (familiar vs novel at 24 h, 3 days, 7 days).

Social isolation selectively disrupts 24 h memory

The paper reports intact short-term memory at 30 min under both grouped and isolated housing, but impaired long-term memory at 24 h after isolation.

Hippocampal lesions dissociate immediate vs delayed memory

The paper reports hippocampal lesion impairment of 30-min recognition but intact immediate post-training recognition, with novel controls indicating specificity of familiar-juvenile decreases.

Protein synthesis inhibition: short-term intact, 24 h impaired

The paper states anisomycin given 30 min before training spares 30-min recognition but disrupts 24-h recognition; Table 2 supports the shift of the anisomycin group toward ~0.5 at 24 h.

CREB αΔ− mutants: short-term intact, long-term impaired

The paper reports that CREB αΔ− mutants show impaired recognition at 24 h but intact 30-min recognition.

Compact table: key recognition-ratio points used above

Paradigm	Condition (familiar)	Test timing	Recognition ratio (mean ± SEM)	n
Baseline (group-housed)	Familiar	24 h	0.41 ± 0.02	12
Baseline (group-housed)	Familiar	3 days	0.39 ± 0.02	10
Baseline (group-housed)	Familiar	7 days	0.40 ± 0.02	10
Isolation	Chronically isolated	30 min	0.37 ± 0.02	16
Isolation	Chronically isolated	24 h	0.52 ± 0.01	15
Isolation	Group-housed	24 h	0.30 ± 0.03	11
Hippocampal lesions	Lesion	30 min	0.48 ± 0.03	8
Hippocampal lesions	Sham	30 min	0.40 ± 0.03	10
Protein synthesis (anisomycin)	Anisomycin	24 h	0.53 ± 0.03	11
Protein synthesis (anisomycin)	Saline	24 h	0.45 ± 0.02	10
CREB αΔ−	CREB αΔ−	24 h	0.51 ± 0.02	7
CREB αΔ−	WT	24 h	0.39 ± 0.03	7

All recognition-ratio values shown above are from Table 2 in the paper.

Skeptical interpretation: what is supported vs what is still uncertain

Supported by multiple converging manipulations (high confidence)

Long-term social recognition exists for group-housed adult male mice for at least 1–7 days after a single 2-min interaction (as defined by the recognition ratio and familiar-vs-novel control logic).
Social isolation disrupts the 24 h memory component, while leaving a 30-min component intact.
Hippocampal lesions and molecular consolidation manipulations (protein synthesis inhibition; CREB αΔ− mutation) alter the long-term component, supporting the paper’s claim that long-term social recognition shares properties with other hippocampus-dependent memory forms.

Uncertainty / potential confounds (moderate confidence)

Olfaction-based encoding is central in rodent social recognition. The paper discusses olfactory cues (including urine/MHC-related signatures). Even if the behavior is “social,” hippocampal dependence may reflect olfactory memory consolidation rather than an abstract social-memory mechanism.
Lesions are not perfectly selective: ibotenic acid lesions can include collateral damage; the paper notes limited neocortical damage above the hippocampus and possible subiculum damage in some cases.
Interpretation of “recognition” vs “general habituation” relies on the novel-juvenile control showing lack of reduced investigation. This control is included, but investigation-time measures can still be sensitive to state changes (arousal/anxiety) induced by manipulations.

What would most likely disprove or revise the paper’s mechanistic framing?

If future work shows that the same behavioral changes can be produced by manipulations that impair odor learning generally (without social-memory-specific consolidation), the “social recognition as a hippocampal memory type” might need reframing.
If isolating mice alters motivation/anxiety or investigation propensity in a way that interacts with timing, the recognition-ratio interpretation could be confounded.
If CREB αΔ− mutation or anisomycin produces broad deficits in performance/states at 24 h (rather than selective consolidation), the claim of CREB-dependent consolidation of social memory would weaken.

Directed “what to check next” (for replication rigor)

Power / sample sizes: some n values (e.g., certain lesion/immediate comparisons, CREB at 24 h) are modest; confidence intervals on the recognition ratio and effect sizes should be checked across repeats.
Lesion verification: quantify lesion extent more finely than “cell loss absence/width reduction,” and test whether lesion extent correlates with 30-min vs immediate vs 24-h performance.
State controls: add explicit measures of arousal/anxiety and baseline olfactory investigation propensity timed to the drug/isolation manipulations, to ensure the ratio is not reflecting altered exploration state.
Stimulus generality: the paper uses juvenile conspecifics from same background and sometimes same litter; testing with matched but distinct juveniles would probe whether the “memory” generalizes or is cue-specific.

Explore related BGPT author reviews

Feedback:

Updated: April 15, 2026