Object recognition memory is the ability to discriminate the familiarity of previously encountered objects. One of our primary research interests has been the study of the neural mechanisms involved in object recognition. We have investigated this issue through the use of delay-non-matching-to-sample (DNMS) tasks and novel object preference (NOP) tests.

This test of object recognition is based on rats' natural propensity to explore novel objects. A rat is allowed to explore two identical sample objects in an open-field arena for a designated amount of time. After a delay, the rat is returned to the arena with two new objects—one is identical to the sample and the other is novel. Normal rats spend more time exploring the novel object, indicating that it recognizes the sample object. We have made considerable use of this novelty-preference test during the past years.

Anterograde Object Recognition after Long Retention Intervals

Fornix lesions have negligible effects on rats' performance on the NOP test, and we have found that cytotoxic (NMDA) lesions of the HPC also fail to disrupt performance with retention delays as long as 24 hours. Others proposed that HPC lesions would impair object recognition after longer delays, but at first, this hypothesis could not be tested, because there had to be a way to demonstrate that normal rats could recognize objects after delays of days or weeks. Neither the DNMS task, nor the standard version of the novelty-preference test can accomplish this. Our solution was to give rats multiple, distributed exposures to the sample object (5 minutes per day on 5 consecutive days) and then test them on their novel object preference. With this modification, normal rats can recognize a sample object on the novelty preference test after delays of up to 7 weeks. We have used this procedure to examine the effects of lesions to the HPC and PeRh on anterograde object recognition after long retention delays, lasting days or weeks.

In one experiment, rats with HPC lesions and control rats were familiarized with a sample object in an open field for 5 minutes a day for 5 consecutive days. Following a 24 h, 1 week or 3 week retention interval, rats were placed back in the apparatus for the retention test. The results showed that the HPC rats’ performance across all three delay conditions was not significantly different from that of control rats (Mumby et al., 2005). These findings suggest that HPC damage spares anterograde object recognition in rats, even after retention intervals lasting weeks.

We also assessed the effects of PeRh lesions on anterograde object recognition using the novel-object preference test, with retention intervals lasting 24 h and 3 weeks. The rats were familiarized with a sample object during the learning phase—5 min per day on 5 consecutive days. Control rats showed a significant novel-object preference after both retention intervals, whereas the rats with PeRh lesions displayed a significant preference only after the 24 hour interval. When the learning phase trial was shortened to a single 5 minute session, the PeRh group was impaired in the 24 h condition (Mumby et al., 2007). The results suggested that the disruptive effects of PeRh damage on anterograde object recognition persist over very long postlearning intervals. Also, the results suggest that object recognition impairments following PeRh damage may not be ubiquitous, and that learning conditions play a significant role in determining the subsequent recognition performance in rats with PeRh damage. In addition to telling us about the contributions of the HPC and PeRh to object recognition, these experiments have yielded information about the temporal limits of object-recognition in normal rats.

We have made significant progress in dissociating the contributions of HPC and PeRh to object- and place-memory under a variety of conditions. In several experiments, we found no evidence that damage to the PeRh impairs learning of a water-maze task, whether working-memory or reference-memory procedures are used (Glenn and Mumby, 1998; Mumby and Glenn, 2000). In contrast, PeRh lesions impair anterograde object-recognition, as assessed by the DNMS task (Mumby and Pinel, 1994), and we have reported that PeRh lesions impair anterograde object recognition on the novelty-preference test when only one familiarization session is provided (Mumby et al., 2002).

We followed up with experiments to determine whether the HPC lesions had failed to produce retrograde amnesia in our first study because the shortest learning-lesion interval (1 week) was too long. In one experiment, rats learned two object discrimination problems 72 hrs before surgery and another two problems 1 hr before surgery. Specifically, 72 hrs before surgery, each rat was trained on one object discrimination problem followed by training on a second problem. The same procedure was used for the learning of two new problems 1 hr prior to surgery. However, for one object discrimination problem a new object pair was introduced, whereas for the second problem rats learned the reversal of one of the problems presented to them 72 hrs before surgery (previous S+ became S-). Postsurgery retention was assessed by presenting trials that resembled presurgery trials, with the exception that an extinction procedure was used (reward was not given for any choices); this ensured that intact anterograde abilities could not obscure the status of retrograde memory. There was no evidence of retrograde amnesia for any of the discrimination problems. Rats with HPC or sham lesions showed a similar bias for selecting the object that was most recently S+, indicating no retrograde amnesia for object-reward associations (temporally-graded amnesia would produce a bias to select the object that was S+ during the remote time period; ungraded amnesia would produce no bias) (Lehmann et al., 2007).

Performance on the object-discrimination task is probably maintained by processes similar to those underlying certain types of learning that are spared in amnesic patients, and it is, therefore, inappropriate for modeling their memory impairment. Tests of recognition are more appropriate. We are now using the procedure of giving multiple exposures to the sample object in the novelty-preference task (5 min on 5 consecutive days) to conduct studies of retrograde object-recognition in brain-damaged rats. In one experiment rats were familiarized with a sample object either 5 weeks or 1 week before receiving HPC lesions. On postsurgery novelty-preference tests, sham-lesion rats recognized the sample objects but HPC rats did not; there was no evidence of a temporal gradient (Gaskin et al., 2003). Given the substantial evidence that HPC damage has negligible effects on anterograde object recognition, we were surprised to find retrograde recognition deficits. Of theoretical significance, the ungraded effect is consistent with the multiple-trace theory of memory consolidation, which gives the HPC a lasting role in long-term memory for certain kinds of information. It is inconsistent with theories which suggest the HPC has only a temporary role in long-term memory. The multiple-trace theory also predicts that the slope of temporal gradients in retrograde amnesia depends on the size of the HPC lesion, with smaller lesions producing steeper temporal gradients, and complete lesions producing severe and ungraded retrograde amnesia. The HPC lesions in all of our studies have been extensive, involving in most cases the entire dorsal hippocampus, and most of the ventral HPC, with variable damage to the subiculum. We plan to examine the effects of lesion size (dorsal HPC vs. ventral HPC vs. complete HPC) on retrograde memory in the water maze and novelty-preference task—the two tasks on which we have already seen non-graded retrograde amnesia after complete HPC lesions.

Retrograde Memory after Extensive Perirhinal Cortex Damage

In several experiments using the object-discrimination task, we found no evidence that Perirhinal cortex lesions impair retrograde memory for object-reward associations. In one experiment, rats learned several object-discrimination problems 4 weeks and 1 week before receiving PeRh lesions. Following recovery, the rats were tested on the previously learned object-discrimination problems. Rats with PeRh lesions displayed no evidence of retrograde amnesia for object-discrimination problems learned either 4 weeks or 1 week before surgery (Mumby et al., 2002).

We also conducted the first studies of retrograde object recognition in rats with PeRh lesions. Rats were familiarized with a sample object, 5 weeks, 3 weeks, or 1 week before receiving PeRh lesions or sham surgery. After surgery, control rats recognized the sample objects but the PeRh rats did not. Importantly, the same rats that received PeRh-lesions displayed normal retrograde memory for object discriminations that they had learned before surgery (Mumby et al., 2002). The findings suggest PeRh damage impairs rats' ability to discriminate the familiarity of objects previously encountered before or after surgery, and that the object-discrimination task does not require this ability.