Determinants of Olfactory Memory Span in Rats MacQueen, D. A., Bullard, L. A., & Galizio, M.

1. Determinants of Olfactory Memory Span in Rats MacQueen, D. A., Bullard, L. A., & Galizio, M. University of North Carolina Wilmington Contact information: galizio@uncw.edu • Introduction • Dudchenko, Wood & Eichenbaum (2000) developed a task that purported to measure olfactory memory span in rats. Their procedure can be described as an incrementing delayed non-matching-to-sample task. On the first trial of a session, digging in a cup of scented sand was reinforced with a food pellet. On Trial 2, a second odor was added to the array and responding to the new odor was reinforced. Each subsequent trial added another odor and the novel odor was always designated correct. • Dudchenko et al. viewed this procedure as analogous to the digit span task used with humans and they reported that rat’s spans (the number of odors that could be identified without error) indicated limitations on working memory similar to those in humans. • The present study replicated the Dudchenko et al. procedures in a large, 18-food cup arena in order to determine whether stable spans could be obtained and to assess the utility of the procedure for behavioral pharmacology. • Methods • Phase I – Initial Training • Six male Sprague-Dawley rats were trained to remove the lids of plastic cups to obtain sucrose pellets. Plastic lids were scented with one of 24 odors by storing them in containers partly filled with household spices (see spice list below). • Subjects were then trained on a non-match to sample task. On Trial 1 the animal was presented with a cup of sand scented with a particular spice baited with once sucrose pellet. On Trial 2 a second cup scented with a different spice was introduced and selection of the novel scent was reinforced. As the session continued, one new odor was added each time the animal made a correct response. • The session was terminated following the first error and the number of correct consecutive trials in a session (excluding the first) defined the span for that session. • Phase II – 18 Comparison Array • Experimental sessions were conducted as described in Phase I with the exception that the session continued for 24 trials regardless of performance and a correction procedure was implemented. • As the apparatus was designed to hold a maximum of 18 stimuli, randomly chosen exemplars were omitted on each of the last six trials in a session • Phase III – 5 Comparison Array • In this phase, although the number of stimuli continued to increment to 24, only five comparisons were presented on any given trial. • Phase IV – Multiple Component • Six simple discrimination trials were added to each session (after every fourth trial) in which only two scents were presented. One scent was always correct, while the other was always incorrect. These two scents were never used in the incrementing NMTS procedure. • Phase V – Scent Marking Component • As Phase IV, but all lids were replaced with fresh lids after each trial. Thus, each lid was used only once per session. Figure 3 shows percent correct for the most recent five sessions for the six rats in scent marking probes. Sessions are broken down into six successive blocks of four trials and percent correct is reported per block. There is a significant effect of trial blocks, F (5, 25) = 5.40 p < .05 which supports the claim that performance declined within the session as the number of stimuli to be remembered incremented. • Conclusions • Five of the six rats developed accurate performances on the incrementing non-match-to-sample task with mean spans ranging from 1.4 to 17.8 by the end of training. • Introduction of the simple discrimination reduced average span which indicates that span is affected by some aspects of the complexity of schedule demands. Performance on the simple discrimination was nearly perfect for all rats throughout the study. • High levels of accuracy persisted even after the initial error of a session. This indicates an important difference between the spans measured here and human performance on the digit span task where accuracy deteriorates dramatically when the span is reached. • However, performance on the NMTS procedure was affected by the number of stimuli to be remembered. Accuracy declined as the memory set incremented. Figure 1.The top panel shows the mean span (number of correct identifications until the first error minus 1) for all six rats during the last five sessions of each phase of training. Note that spans in Phase 2 were increased substantially relative to those obtained in Phase 1, presumably due to continued training with the procedure. Moving to the 5-comparison array slightly reduced spans and spans declined further when the simple discrimination was introduced in Phase 4. The lower panel shows percent correct for the same sessions of Phase 2-5. Note that percent correct remained high regardless of span, indicating that accurate performances were maintained even after an initial error occurred. Also, percent correct remained high in the final phase which used fresh stimuli in each trial to prevent scent marking. Apparatus Figure 2 shows the mean span and number of correct trials (out of 24) for individual subjects during the last five sessions of the scent marking component. While the average span is consistent with the findings of Dudchenko et al., It is important to note that within subject variability is much greater for the span measure. SPICE LIST 1 – Fennel 2 – Clove 3 – Thyme 4 – Beet 5 – Onion 6 – Turmeric 7 – Paprika 8 – Caraway 9 – Coriander 10 – Cinnamon 11 – Bay 12 – Ginger 13 – Nutmeg 14 – Savory 15 – Garlic 16 – Mustard 17 – Sage 18 – Sumac 19 – Cumin 20 – Oregano 21 – Marjoram 22 – Dill 23 – Celery 24 - Rosemary Reference: Dudchenko PA, Wood ER, Eichenbaum H. (2000). Neurotoxic hippocampal lesions have no effect on odor span and little effect on odor recognition memory but produce significant impairments on spatial span, recognition, and alternation. J Neurosci 20, 2964-77.