Figure 2. Memory impairments emerge latently, weeks after the symptoms of transient illness subside.

LPS- (5mg/kg; black symbols and bars) and saline- (white symbols and bars) injected mice were trained on hidden platform trials in the Morris water maze beginning on Day 7 (n = 8 mice per group included in analysis) or Day 28 (n = 9 saline- and 10 LPS-injected mice included in analysis) after injection. (A) Schematic showing the shorter pathlengths (m) and latencies (s) to the hidden platform across sessions that are indicative of spatial learning. (B–D) Transient illness does not spatial learning or memory in mice trained on Day 7 after LPS. LPS- and saline-injected mice exhibited similar pathlengths (interaction effect: p = 0.75) across visible platform trials administered on Day 6 (B) and across hidden platform training sessions initiated on Day 7 (interaction effect: p = 0.732; C). All mice swam shorter pathlengths on later versus earlier hidden platform training sessions (Session 1 > 2–8; p values < 0.001). Both groups preferred the target versus the left (p = 0.07), right (p = 0.002) or opposite (p = 0.0003) quadrant in the memory probe trial administered one week after training (D). (E and -F) Mice exhibit intact ability to learn a hidden platform location one month after transient illness resolves but impaired memory. All mice swam shorter pathlengths on later versus earlier trials (Sessions 1, 2 > 3 ≥ 4, 5 > 8–10, p values < 0.02). The interaction effect (p = 0.404) confirmed that the group differences on Sessions 7, 9 and 10 were not statistically significant (E). Although all mice that learned the hidden platform location preferred the training versus left, right or opposite quadrants (p values < 0.0002) on a the probe trial administered one week after training, LPS-injected mice spent significantly less time in the training quadrant (*p = 0.042) than controls.