Human working memory is a capacity- and duration-limited system in which retention and manipulation of information is subject to metacognitive monitoring and control. At least some nonhuman animals appear to also monitor and control the contents of working memory, but only relatively simple cases where animals monitor or control the presence or absence of single memories have been studied. Here we combine a comparatively complex order memory task with methodology that assesses the capacity to introspect about memory. Monkeys observed sequential presentations of five images, and at test, reported which of two images from the list had appeared first during study. Concurrently, they chose to complete or avoid these tests on a trial-by-trial basis. Monkeys “knew when they knew” the correct response. They were less accurate discriminating images that had appeared close in time to one another during study and were more likely to avoid these difficult tests than they were to avoid easier tests. These results indicate that monkeys can metacognitively monitor relatively complex properties of the contents of working memory, including the quality of representations of temporal relations among images.
Adaptive decision making in humans depends on feedback between monitoring, which assesses mental states, and control, by which cognitive processes are modified. We investigated the extent to which monitoring and control interact iteratively in monkeys. Monkeys classified images as birds, fish, flowers, or people. At the beginning of each trial, to-be-classified images were not visible. Monkeys touched the image area to incrementally brighten the image, referred to as the brighten response. The amount by which brightness increased with each brighten response was unpredictable, and the monkeys could choose to classify the images at any time during a trial. We hypothesized that if monkeys monitored the status of their classification decision then they would seek information depending on the amount of information available. In Experiment 1, monkeys rarely used the brighten response when images were bright initially, and they used the brighten response more when earlier uses in a given trial yielded smaller amounts of information. In Experiment 2, monkeys made more brighten responses when the presented image did not belong in any of the trained categories, suggesting monkeys were sensitive to the fact that they could not reach a classification decision despite the image brightening. In Experiment 3, we found that the probability that monkeys used the brighten response correlated with their ability to correctly classify when the brighten response was not available. These findings add to the literature documenting the metacognitive skills of nonhuman primates by demonstrating an iterative feedback loop between cognitive monitoring and cognitive control that allows for adaptive information-seeking behavior.
Monkeys with selective damage to the hippocampus are often unimpaired in matching-to-sample tests but are reportedly impaired in visual paired comparison. While both tests assess recognition of previously seen images, delayed matching-to-sample may engage active memory maintenance whereas visual paired comparison may not. Passive memory tests that are not rewarded with food and that do not require extensive training may provide more sensitive measures of hippocampal function. To test this hypothesis, we assessed memory in monkeys with hippocampal damage and matched controls by providing them the opportunity to repeatedly view small sets of videos. Monkeys pressed a button to play each video. The same 10 videos were used for six consecutive days, after which 10 new videos were introduced in each of seven cycles of testing. Our measure of memory was the extent to which monkeys habituated with repeated presentations, watching fewer videos per session over time. Monkeys with hippocampal lesions habituated more slowly than did control monkeys, indicating poorer memory for previous viewings. Both groups dishabituated each time new videos were introduced. These results, like those from preferential viewing, suggest that the hippocampus may be especially important for memory of incidentally encoded events.
Animals housed in naturalistic social groups with access to automated cognitive testing vary in whether and how much they participate in cognitive testing. Understanding how demographic, seasonal, and social factors relate to participation is essential to evaluating the usefulness of these systems for studying cognition and in assessing the data produced. We evaluated how sex, age, reproductive experience, seasonality, and rank related to patterns of participation in a naturalistic group of rhesus monkeys over a 4-year period. Females interacted with the touchscreen systems more than males and were more likely to complete initial training. Age was positively correlated with touchscreen activity through adolescence in females, at which point seasonality and reproductive experience were stronger associates of participation. While monkeys in different rank categories did not differ in how much they interacted with the touchscreen systems, monkeys of different ranks tended not to work at the same times, perhaps reflecting avoidance of high ranking animals by those of lower rank. Automated cognitive testing systems for naturalistic social groups of rhesus monkeys can yield quality cognitive data from individuals of all ages and ranks, but participation biases may make it difficult to study sex differences or seasonal variation in cognition.
Working memory is a system by which a limited amount of information can be kept available for processing after the cessation of sensory input. Because working memory resources are limited, it is adaptive to focus processing on the most relevant information. We used a retro-cue paradigm to determine the extent to which monkey working memory possesses control mechanisms that focus processing on the most relevant representations. Monkeys saw a sample array of images, and shortly after the array disappeared, they were visually cued to a location that had been occupied by one of the sample images. The cue indicated which image should be remembered for the upcoming recognition test. By determining whether the monkeys were more accurate and quicker to respond to cued images compared to un-cued images, we tested the hypothesis that monkey working memory focuses processing on relevant information. We found a memory benefit for the cued image in terms of accuracy and retrieval speed with a memory load of two images. With a memory load of three images, we found a benefit in retrieval speed but only after shortening the onset latency of the retro-cue. Our results demonstrate previously unknown flexibility in the cognitive control of memory in monkeys, suggesting that control mechanisms in working memory likely evolved in a common ancestor of humans and monkeys more than 32 million years ago. Future work should be aimed at understanding the interaction between memory load and the ability to control memory resources, and the role of working memory control in generating differences in cognitive capacity among primates.
Evidence that the hippocampus is critical for spatial memory in nonnavigational tests is mixed. A recent study reported that temporary hippocampal inactivation impaired spatial memory in the nonnavigational Hamilton Search Task in monkeys. However, several studies have documented no impairment on other nonnavigational spatial memory tests following permanent hippocampal lesions. It was hypothesized that transient, but not permanent, hippocampal disruption produces deficits because monkeys undergoing transient inactivation continue to try to use a hippocampal-dependent strategy, whereas monkeys with permanent lesions use a nonhippocampal-dependent strategy. We evaluated this hypothesis by testing five rhesus monkeys with hippocampal lesions and five controls on a computerized analogue of the Hamilton Search Task.
On each trial, monkeys saw an array of squares on a touchscreen, each of which “hid” one reward. Retrieving a reward depleted that location and monkeys continued selecting squares until they found all rewards. The optimal strategy is to remember chosen locations and choose each square once. Unlike the inactivation study, monkeys with hippocampal damage were as accurate as controls regardless of retention interval. Critically, we found no evidence that the groups used different strategies, as measured by learning rates, spatial search biases, perseverative win-stay errors, or inter-choice distance. This discrepancy between the effect of inactivations and lesions may result from off-target effects of inactivations or as-yet-unidentified differences between the physical and computerized tasks.
Combined with previous evidence that hippocampal damage impairs navigational memory in monkeys, this evidence constrains the role of the hippocampus in spatial memory as being critical for navigational tests that likely involve allocentric spatial memory but not nonnavigational tests that likely involve egocentric spatial memory.
Human working memory is greatly facilitated by linguistic representations—for example, by verbal rehearsal and by verbal recoding of novel stimuli. The absence of language in nonhumans raises questions about the extent to which nonhuman working memory includes similar mechanisms. There is strong evidence for rehearsal-like active maintenance in working memory when monkeys are tested with highly familiar stimuli, but not when tested with novel stimuli, suggesting that working memory depends on the existence of previously encoded representations. This difference in working memory for familiar and novel images may exist because, lacking language, monkeys cannot recode novel stimuli in a way that permits active maintenance in working memory.
Alternatively, working memory for novel images may have been present, but behaviorally silent, in earlier studies. In tests with novel images, the high familiarity of to-be-remembered stimuli compared to never-before-seen distractors may be such a strong determinant of recognition performance that evidence of working memory is obscured. In the current study, we developed a technique for attenuating the utility of relative familiarity as a mnemonic signal in recognition tests with novel stimuli. In tests with novel images, we observed impairments of memory by concurrent cognitive load and delay interval that indicate actively maintained working memory. This flexibility in monkey working memory suggests that monkeys may recode unfamiliar stimuli to facilitate working memory and establishes new parallels between verbal human working memory and nonverbal nonhuman primate working memory.
Some nonhuman species demonstrate metamemory, the ability to monitor and control memory. Here, we identify memory signals that control metamemory judgments in rhesus monkeys by directly comparing performance in two metamemory paradigms while holding the availability of one memory signal constant and manipulating another. Monkeys performed a four-choice match-to-sample memory task. In Experiment 1, monkeys could decline memory tests on some trials for a small, guaranteed reward. In Experiment 2, monkeys could review the sample on some trials. In both experiments, monkeys improved accuracy by selectively declining tests or reviewing samples when memory was poor. To assess the degree to which different memory signals made independent contributions to the metamemory judgement, we made the decline-test or review-sample response available either prospectively, before the test, or concurrently with test stimuli.
Prospective metamemory judgements are likely controlled by the current contents of working memory, whereas concurrent metamemory judgements may also be controlled by additional relative familiarity signals evoked by the sight of the test stimuli. In both paradigms, metacognitive responding enhanced accuracy more on concurrent than on prospective tests, suggesting additive contributions of working memory and stimulus-evoked familiarity. Consistent with the hypothesis that working memory and stimulus-evoked familiarity both control metamemory judgments when available, metacognitive choice latencies were longer in the concurrent condition, when both were available. Together, these data demonstrate that multiple memory signals can additively control metacognitive judgements in monkeys and provide a framework for mapping the interaction of explicit memory signals in primate memory.
Many species classify images according to visual attributes. In pigeons, local features may disproportionately control classification, whereas in primates global features may exert greater control. In the absence of explicitly comparative studies, in which different species are tested with the same stimuli under similar conditions, it is not possible to determine how much of the variation in the control of classification is due to species differences and how much is due to differences in the stimuli, training, or testing conditions. We tested rhesus monkeys (Macaca mulatta) and orangutans (Pongo pygmaeus and Pongo abelii) in identical tests in which images were modified to determine which stimulus features controlled classification. Monkeys and orangutans were trained to classify full color images of birds, fish, flowers, and people; they were later given generalization tests in which images were novel, black and white, black and white line drawings, or scrambled. Classification in these primate species was controlled by multiple stimulus attributes, both global and local, and the species behaved similarly.
Dissociations of memory systems are typically made using independent cognitive tests. For example, in monkeys habits have been inferred from performance in object discrimination tests, while nonmatching-to-sample tests are thought to measure familiarity resulting from single exposures. Such tests cannot measure individual memory processes accurately when more than one memory process contributes to performance. In process dissociation procedures (PDPs) two memory processes cooperate and compete in the performance of a single cognitive task, allowing quantitative estimates of the contributions of each process. We used PDP to measure the contributions of habits and one-trial memory to visual matching-to-sample performance. Sets of test images were shown only once in each daily testing session but were repeated day after day. To produce habits, high-frequency images were correct more frequently than other images across days. Habits were manifest in the extent to which choices in the test phase of matching-to-sample trials were made to the high-frequency images, irrespective of which image had been presented as the sample. One-trial memory was measured by the extent to which choices at test were made to the image that had appeared as the sample on that trial, irrespective of habit. Perirhinal cortex removal reduced the contribution of one-trial memory to matching performance, but left both habits and the ability to discriminate images intact. PDP can be applied in monkeys in a way that parallels its use in humans, providing a new tool for investigating the neurobiology of memory in nonhuman animals and for comparing memory across species.