Unsupervised, frequent and remote: A novel platform for personalised digital phenotyping of long-term memory in humans

In a continuous T-maze spatial alternation task, CA1 place cells fire differentially on the stem of the maze as rats are performing left- and right-turn trials (Wood et al. (2000) Neuron 27:623-633). This context-dependent hippocampal activity provides a potential mechanism by which animals could solve the alternation task, as it provides a cue that could prime the appropriate goal choice. The aim of this study was to examine the relationship between context-dependent hippocampal activity and spatial alternation behavior. We report that rats with complete lesions of the hippocampus learn and perform the spatial alternation task as well as controls if there is no delay between trials, suggesting that the observed context-dependent hippocampal activity does not mediate alternation behavior in this task. However lesioned rats are significantly impaired when delays of 2 or 10 s are interposed. Recording experiments reveal that context-dependent hippocampal activity occurs in both the delay and no-delay versions of the task, but that in the delay version it occurs during the delay period, and not on the stem of the maze. These data are consistent with a role for context-dependent hippocampal activity in delayed spatial alternation, but suggest that, according to specific task demands and memory load, the activity may be generated by different mechanisms and/or in different brain structures.

Effects of treadmill exercise intensity on spatial working memory and long-term memory in rats

Polychlorinated biphenyls (PCBs) and methylmercury (MeHg) are ubiquitous environmental contaminants that alter cognitive function in both humans and animals. Because PCBs and MeHg often occur together in the environment, it is important to understand whether these two contaminants have the potential to interact, causing additive or greater than additive effects. The current study examined the combined effects of gestational and lactational exposure to Aroclor 1254 (A1254), a commercial PCB mixture, and MeHg on a series of spatial alternation tasks including cued spatial alternation (CA), non-cued spatial alternation (NCA), and delayed spatial alternation (DSA) in rats using standard two-lever operant testing chambers. Pregnant Long-Evans rats received either 6 mg/kg A1254 pipetted onto a Keebler Vanilla Wafer cookie (PCB-only group), 0.5 ppm. MeHg dissolved in the drinking water (MeHg-only group), 6 mg/kg A1254 + 0.5 ppm. MeHg (PCB + MeHg group), or corn oil vehicle and normal tap water (control group) beginning 28 days prior to mating and continuing through postnatal day 16. One male and one female from each litter began testing on spatial alternation at approximately 110 days of age. Animals were reinforced for pressing the lever opposite that pressed on the previous trial. In general, animals exposed to A1254 and/or MeHg were impaired relative to control rats on the NCA and DSA tasks. Significant reductions in NCA performance were observed in the MeHg-only and PCB + MeHg groups, while significant reductions in DSA performance were observed in the PCB-only and MeHg-only groups. The PCB + MeHg group showed a similar magnitude reduction in performance on DSA, but this difference was not statistically significant due to increased variability in that group. The reductions in DSA performance were observed across most of the delays, indicating that memory impairments were not likely the cause of the deficit. Instead, the DSA deficits following exposure to A1254 and/or MeHg are indicative of either an associative or attentional impairment. The results from the current study indicate that combined exposure to PCBs and MeHg does not exacerbate the PCB- or MeHg-induced impairments on spatial alternation tasks.

Animal behavior provides context for understanding disease models and physiology. However, that behavior is often characterized subjectively, creating opportunity for misinterpretation and misunderstanding. For example, spatial alternation tasks are treated as paradigmatic tools for examining memory; however, that link is actually an assumption. To test this assumption, we simulated a reinforcement learning (RL) agent equipped with a perfect memory process. We found that it learns a simple spatial alternation task more slowly and makes different errors than a group of male rats, illustrating that memory alone may not be sufficient to capture the behavior. We demonstrate that incorporating spatial biases permits rapid learning and enables the model to fit rodent behavior accurately. Our results suggest that even simple spatial alternation behaviors reflect multiple cognitive processes that need to be taken into account when studying animal behavior.SIGNIFICANCE STATEMENT Memory is a critical function for cognition whose impairment has significant clinical consequences. Experimental systems aimed at testing various sorts of memory are therefore also central. However, experimental designs to test memory are typically based on intuition about the underlying processes. We tested this using a popular behavioral paradigm: a spatial alternation task. Using behavioral modeling, we show that the straightforward intuition that these tasks just probe spatial memory fails to account for the speed at which rats learn or the types of errors they make. Only when memory-independent dynamic spatial preferences are added can the model learn like the rats. This highlights the importance of respecting the complexity of animal behavior to interpret neural function and validate disease models.

The retrosplenial cortex (RSC) plays an important role in spatial cognition. RSC neurons exhibit a variety of spatial firing patterns and lesion studies have found that the RSC is necessary for spatial working memory tasks. However, little is known about how RSC neurons might encode spatial memory during a delay period. In the present study, we trained control rats and rats with excitotoxic lesions of the RSC on spatial alternation task with varying delay durations and in a separate group of rats, we recorded RSC neuronal activity as the rats performed the alternation task. We found that RSC lesions significantly impaired alternation performance, particularly at the longest delay duration. We also found that RSC neurons exhibited reliably different firing patterns throughout the delay periods preceding left and right trials, consistent with a working memory signal. These differential firing patterns were absent during the delay periods preceding errors. We also found that many RSC neurons exhibit a large spike in firing rate leading up to the start of the trial. Many of these trial start responses also differentiated left and right trials, suggesting that they could play a role in priming the ‘go left’ or ‘go right’ behavioral responses. Our results suggest that these firing patterns represent critical memory information that underlies the RSC role in spatial working memory.

The retrosplenial cortical role in delayed spatial alternation

Human cognitive processes are highly variable across individuals and are influenced by both genetic and environmental factors. Although genetic variations affect short-term memory in humans, it is unknown whether genetic variability has also an impact on long-term memory. Because prion-like conformational changes may be involved in the induction of long-lasting synaptic plasticity, we examined the impact of single-nucleotide polymorphisms (SNPs) of the prion protein gene (PRNP) on long-term memory in healthy young humans. SNPs in the genomic region of PRNP were associated with better long-term memory performance in two independent populations with different educational background. Among the examined PRNP SNPs, the common Met129Val polymorphism yielded the highest effect size. Twenty-four hours after a word list-learning task, carriers of either the 129MM or the 129MV genotype recalled 17% more information than 129VV carriers, but short-term memory was unaffected. These results suggest a role for the prion protein in the formation of long-term memory in humans.

Y-mazes are widely used for discrimination learning and spatial alternation tasks. We present here a computer-assisted Y-maze model for spatial alternation learning via footshock reinforcement. No intramaze cues are provided and handling between trials is no longer required. Pretraining application of scopolamine (1 mg kg-1) and D-AP5 (2-amino-5-phosphonopentanoate; 0.016 mg) clearly impairs acquisition and retention (tested 24 h after acquisition) of spatial alternation. These effects are not due to state-dependent changes caused by drug treatment. Similar results were reported by others in food-motivated spatial alternation tasks as well as spontaneous alternation paradigms. The data support our model as a useful tool for studying spatial alternation in the Y-maze.

Alternation behavior of cats with medial visual cortex ablation

Dissociable roles of the dorsal striatum and dorsal hippocampus in conditional discrimination and spatial alternation T-maze tasks

Role of residual anterior neocortex in recovery from neonatal prefrontal lesions in the rat

We examined the effects of both the metabotropic glutamate receptor (mGluR) antagonist MCPG and the agonist tADA in two behavioral paradigms in rats: (1) brightness discrimination and (2) spatial alternation. Compounds were applied intracerebroventricularly at different times, either 30 min prior to training or immediately after training, and rats were tested for retention 24 hr later in the same paradigms. Both MCPG and tADA caused amnesia in the spatial alternation test, when applied pretraining, but no effect was obtained in the brightness discrimination paradigm. Drug-induced amnesia was shown not to be attributable to state-dependent effects of MCPG or tADA. Moreover, the memory inhibiting effect of MCPG was dose dependent, with a low dose (20 mM/5 ml) having no effect on learning and memory and a 10 times higher concentration (200 mM/5 ml) causing complete amnesia. Application of both saline and MCPG immediately post-training prevented memory formation, which may be attributable to an interference by the injection procedure with the process of memory formation. The mGluR agonist tADA, however, facilitated memory formation in the spatial alternation task, when injected immediately after training. Post-training application of the compounds had no effect on retention in the brightness discrimination task. On the basis of these data we conclude that (1) mGluRs are of particular importance for spatial learning and play no role in visual discrimination; (2) both the block and the activation of mGluRs inhibit spatial learning, suggesting that saturated activation prevents further modulation of mGluRs, which may be required during learning or memory formation; and (3) mGluR agonist tADA may be memory facilitating when applied after training, thus enhancing the establishment of the memory trace.

Converging behavioral, electrophysiological, and neurochemical data suggest that lesions to the peripheral vestibular system result in impairment of the hippocampus. Nonetheless, relatively few studies have examined the hippocampus or behavior related to it, over a long period of time following the lesion, to determine if any recovery takes place. Here we used the spatial forced alternation task in a T maze, which is sensitive to the integrity of the hippocampus, to evaluate learning and memory in rats at 3 weeks, 3 months, and 5 months following bilateral vestibular deafferentation (BVD) or sham surgery. BVD rats made significantly fewer correct choices at all time points when compared with the sham controls. However, the percentage correct choice for BVD rats was at chance level at 3 weeks postop, and was significantly above chance at 5 months postop. These results add to the evidence that BVD causes a long-term impairment of hippocampal function and spatial learning and memory, but suggest that some recovery of function might take place over the long term.

Neurotoxic lesions of the amygdala did not affect the postoperative acquisition of a nonspatial test of object recognition (delayed nonmatching to sample) even when retention delays were increased from 0 s to 20 or 60 s, or when test stimuli were deliberately repeated within a session. Although these amygdaloid lesions did not alter forced-choice spatial alternation, they slightly increased neophobic responses to novel foods and environments. In contrast, combined amygdalohippocampal (A + H) lesions impaired performance on the object recognition task when the retention intervals were increased beyond 0 s and when test stimuli were repeated within a session. The A + H rats were also severely impaired on the spatial alternation task, and they showed reduced neophobia. Comparisons with a previous study show that damage to the amygdala or hippocampus does not affect object recognition, whereas A + H damage produces clear deficits.

Rats, one groups with fornicotomies, and another group with control operations, were trained in several versions of a bar-press, spatial alternation task. In free operant conditions the fornicotomized rats showed adequate alternation performance when the bars were relatively separated by placement either at the ends of an alley or at the ends of the arms of a T-maze, but they were impaired when the bars were adjacent. Also, the fornicotomized rats failed to alternate after interpolation of either baffles or a 10-sec delay in the stem, both of which manipulations resulted in intervening turning responses before each choice. A failure to discriminate memories of relevant, discrete events from those of similar, intervening events appears best to account for the results.