Behavioral tests of cognition have had a long tradition in the field of primate neuropsychology. Originally, testing took place in the Wisconsin General Test Apparatus
(WGTA) with the experimenter sitting in front of the monkey, behind a one-way mirror. Given that vision is an extremely important sense for monkeys, tests tended to be designed around objects or spatial locations. This contrasted with that of human neuropsychology in which the tests so often involved language and pen and paper. Consequently, the extrapolation of results from primate studies into the clinic was fraught with difficulty, as the tests used to measure a particular cognitive process differed considerably between humans and monkeys. After a seminal paper by L Weiskrantz, in 1977, the recognition of the advantages of testing humans and monkeys on the same tests led to much closer integration of human and monkey neuropsychological studies. In addition, the important advantages of automated and computer controlled testing devices over manually operated ones (e.g., WGTA) were recognized, eliminating experimenter-subject interactions and increasing the degree of experimental control and efficiency (Bartus and Dean 2009). However, the ease of presenting tests in the WGTA or other manually operated environments should not be underestimated, not least due to the comparative ease at which monkeys can learn a range of different cognitive tests when the response and the reward are spatially contiguous. The spatial (and thus also temporal) separation of a response and the associated reward recruits additional cognitive processes, probably dependent on the frontal lobes, which may not be the focus of interest, but may need to be taken into account when interpreting the results.
Currently, there are an array of cognitive tests designed to measure specific aspects ofprimate cognition, which are available for the psychopharmacologist. In many cases, it is possible to test monkeys on a battery of such tests allowing the effects of drugs to be compared across a range of different cognitive functions within the same animal.
Attention can be selective or ► divided, ► sustained or not. If selective, the attention may be directed at a specific spatial location or a particular sensory cue e.g., red circle. Alternatively, ► attention may transcend specific sensory cues and occur instead at the level of higher-order perceptual dimensions e.g., color or shape, in which case, it is said that an animal has developed an ► attentional set (see the section on Cognitive Flexibility). A variety of tests have been developed to study attentional abilities in monkeys and the majority of them are dependent on an intact frontal and parietal cortex. The serial reaction time task, first developed in humans and later used to study attention in rats, investigates the ability of monkeys to locate a briefly presented target in one of a number of spatial locations (Spinelli et al. 2004; Weed et al. 1999). It tests aspects of both divided and selective attention. In a version for monkeys (Cambridge Neuropsychological Test Battery, ► CANTAB. Lafayette Instrument company), five circles are presented on a touch sensitive computer screen, and a small colored stimulus is briefly presented in one of those locations (Fig. 1a). To start each trial, the monkey must perform an orienting response to ensure readiness to perform the trial, i.e., press down a lever, and after a variable delay, respond to where they saw the colored stimulus. The demands on divided attention can be increased or decreased by altering the number of spatial locations in which the target stimulus may appear. In contrast, the level of selective attention can be modulated by varying the duration of the target presentation. Other manipulations involve altering the lengths of the inter-trial interval and the length of time the animal waits at the start of a trial for the onset of the target stimulus, both of which increase overall difficulty and reduce successful performance. Such manipulations are especially useful if the cognitive enhancing effects of a drug are under investigation. Another test of attention that assesses the monkey's ability to focus attention using advanced information is a cued reaction time test. Here, four outlines of circles are presented on a touch sensitive computer screen and similar to that described earlier, the monkey must depress a lever to start the trial and respond to the circle that turns white as quickly as possible. In a cued version, a cue light appears above the circle that will become the next target thereby improving the speed of reaction time to that target. By comparing reaction and movement times in the cued and uncued conditions, a specific measure of selective attention can be obtained (Decamp and Schneider 2004). A variation of this task, first developed by MI Posner and colleagues, tests the abilities of monkeys to shift visuospatial attention and includes trials in which the cue is invalid and the target appears on the side opposite to that of the cue. The difference in the speed of responding to a target at expected (valid) and unexpected (invalid) locations is taken as a measure of ability to shift attention.
The classic test of recognition memory or the judgment of the prior occurrence of an object involves monkeys being presented with a novel object and then, after a variable delay (e.g., 5 s to 24 h) being presented with two objects, the previously seen object and a novel object. In the ► delayed "match to sample'' (DMS) version, monkeys have to select the previously seen object while in the ► delayed "nonmatch to sample'' (DNMS) version monkeys have to choose the novel object. It was first described by M Mis-hkin and J Delacour in 1975. This test differs from the working memory tasks described in the following text in that each pair of objects is only seen once, or at least only once within a session and therefore tests the ability of monkeys to recognize objects as familiar or not. The presence of delay-dependent deficits is taken to indicate a mnemonic impairment while poor performance at very short delays might indicate, instead, a perceptual deficit. However, to avoid artifacts attributable to the provision of extensive experience at short delays, followed by testing with less familiar long delays, it is important to intermix different delay lengths. For psychopharmacological studies, it is often beneficial to titrate the duration of the delay interval of each monkey to obtain matched levels of performance accuracy. This helps to equate levels of difficulty across monkeys and to avoid ceiling effects in the highest performing monkeys (Buccafusco 2008). More recently, D(N)MS has been run with computer graphic stimuli presented on touch screen monitors (e.g., Fig. 1b) by a number of different research laboratories including those of D Gaffan and EA Murray. In the DMS version, the number of stimuli at the time of choice can be varied, along with the degree to which the different choices differ from one another perceptually. Drug manipulations during sample presentation, choice phase and during the retention interval can target ► encoding, ► retrieval and ► consolidation, respectively. Two distinct processes may underlie recognition memory, recollection, and familiarity judgment. Performance on this test is dependent upon an intact perirhinal cortex.
A variety of tests have been used to study visuospatial memory, which is the ability to integrate visual and spatial information and to recall that information subsequently. It is dependent on the ► hippocampus and related circuitry. One such test is a scene discriminations task in which, on each trial, an artificially constructed unique "scene" is presented consisting of randomly selected attributes including a colored background containing ellipse segments of different colors, sizes, and orientations and typographical characters (Gaffan and Parker 1996). In the foreground, one of two objects is rewarded. Monkeys learn these discriminations over a series of sessions in which each unique scene is presented once each session. Subsequently it is possible to investigate the ability of monkeys to recall these discriminations, to relearn them as well as to acquire new discriminations and thus, as for recognition memory,
Primate Models of Cognition. Fig. 1. Examples of some of the cognitive tasks used successfully in studies of primate psychopharmacology. a. The five choice serial reaction time (SRT) test used in monkey CANTAB. in which on each trial, after a variable delay, monkeys have to detect the brief presentation of a stimulus in one of five spatial locations. b. An example of a recognition memory test depicting stimuli similar to those used in monkey CANTAB. See text for details. c. A visuospatial learning and memory test also from monkey CANTAB. The stimuli used here are for illustrative purposes and are not the actual stimuli used in monkeys CANTAB. Examples of one, two or three stimulus trial types are illustrated. d. A spatial search task in which monkeys have to respond once, and once only to each of a number of spatial locations on the screen (out of a possible eight locations) in order to get reward. An example of 2,3 and 5 box problems are shown. The different colors help to differentiate the different box problems from one another. ei. Example of a discrimination reversal task, as described in the text. Which stimulus is rewarded and which is not is shown by the "+" and "—'' signs, respectively. An example of a test to separate out whether a reversal deficit is due to perseveration or learned avoidance is shown in eii. f. Examples of a series of discriminations involving an ID shift, probe test, and ED shift, as described in the text.
to investigate encoding and retrieval. An alternative test, known as the paired associate learning test (vsPAL, CANTAB) involves learning to associate different patterns in distinct spatial locations (Fig. 1c). On any one trial (sample phase), one or more patterns may be presented, one at a time, in distinct spatial locations and the monkey must learn which pattern(s) is(are) associated with which spatial location(s). The monkeys are subsequently tested on this knowledge (choice phase) by presenting the pattern in two or more spatial locations and requiring them to respond to the pattern in the correct spatial location (Taffe et al. 2004). Performance on the first attempt of the choice phase is a test of how well monkeys recall the information. However, subsequently monkeys are presented the sample and choice phases, an additional number of times to determine how rapidly they can learn this information. Because on the first attempt, monkeys have to remember multiple information across a short delay, this test loads quite heavily on working memory and is sensitive to frontal lobe dysfunction. This particular test is able to differentiate probands who will subsequently be diagnosed with Alzheimer dementia as opposed to other forms of mnemonic deficit related to normal aging and depression.
Executive functions can be thought of as general-purpose control mechanisms that coordinate specific cognitive processes in order to optimize performance. They include holding information on-line and updating information in working memory, marshaling attentional resources, monitoring behavior and its outcomes, and inhibiting inappropriate strategies and responses.
► Working memory is a short-term memory system that allows for the active maintenance and manipulation of information that is not present in the outside world. This concept of working memory involves temporary buffers of information and an attentionally constrained central executive. Tests of working memory measure a monkey's ability to remember information for a short period of time when that information is no longer present in the environment. Whether the information is remembered or not is determined by how well their response is guided by this information after a variable delay. The information to be remembered is usually visual, i.e., remembering the spatial location of a stimulus or the form/color of a stimulus, but other sensory modalities can be used. It involves presenting a stimulus to the monkey for a brief period of time (seconds) and then after a variable delay requiring the subject to use that information to guide their subsequent responding. One of the most well-known versions of this test is the spatial working memory test, whereby a monkey either watches food reward being hidden in one of two food wells to the left and right of the animal or alternatively fixates a central cross while a spot of light is briefly flashed (e.g., 100 ms) in one of eight spatial locations in the surround. In both cases, the animal must remember the location of the stimulus for a brief delay, e.g., 1-30 s, and then make a response (arm reach/ saccade) to the remembered location at the end of the delay period. In object working memory tests, an object is presented centrally in a WGTA or a visual pattern is presented on a computer screen (sample phase) for a short period of time during which the monkey may have to respond to the stimulus to demonstrate that the stimulus has been seen (Weed et al. 1999). Following a variable delay the same visual stimulus, along with another one is presented to the left and right of the center, and the animal has to choose the object/pattern that matches the sample object/pattern. The smaller the stimulus set used and thus, how frequently the same stimulus is seen across trials, increases the level of interference between trials and increases the load on working memory. By comparing responding following a delay with that at zero delay it is possible to look at the effects of a drug on those processes specifically involved in short term/working memory as compared to other more general perceptual and rule learning processes. In the spatial versions of the test it is important to ensure that the monkey does not use a mediating response to bridge the delay between the stimulus disappearing and the response being made and thus reducing the load on working memory. The addition of distraction during the delay period allows for attentional mechanisms and working memory processes to be assessed within the same session. The distractor can take the form of additional, related stimuli presented during the delay that are irrelevant to the task or alternatively, a burst of loud noise may be presented.
Self-ordered search tasks. Self-ordered search tasks were originally used in studies of human frontal lobe and were later adapted for studies in monkeys by M Petrides. In these original versions, monkeys were presented with variable numbers of objects (usually three) and across a series of three trials they were presented with the same three objects. Food could be retrieved each trial as long as they selected a different object on each occasion. Thus, the monkey must monitor within working memory its earlier choices in order to avoid returning to objects that have already been selected. Spatial versions require animals to search through a series of spatial locations in order to find food reward. In those versions of self-ordered tasks in which food is obtained by selecting an object or spatial location for the first time, but not on repeated selections, the task is very similar to a foraging test. Alternative, computerized versions in which stimuli are presented on a touch sensitive computer screen (Fig. 1d) tend not to reward monkeys until they have responded once and once only to a series of spatial locations (Collins et al. 1998). In the latter, monkeys are having to perform sequences of actions in order to gain reward, which may depend upon distinct prefrontal circuitry to those tasks in which the stimuli themselves are associated with reward. Besides monitoring of actions, the spatial search tasks can also be used to assess simple planning ability, especially if the number of boxes/locations is increased, as implementation of a strategy, such as following a clockwise or anticlockwise strategy reduces demands upon working memory.
Cognitive flexibility is the ability of animals to adapt their responding to changes in the environment. In the tasks described in the following text, previously rewarded responses or strategies have to be inhibited in favor of the development of new responses and strategies. Different aspects of cognitive flexibility are associated with different regions of ► prefrontal cortex, with many neuropsychia-tric disorders being associated with cognitive inflexibility.
Discrimination reversal tasks. These typically involve presenting two stimuli to a monkey, usually two visual objects, and the animal learns, through trial and error, that a response to one of the objects leads to food reward while a response to the other does not. Having learnt this discrimination to a particular level of performance, usually around 90% correct over a series of trials, the reward contingencies reverse such that the previously rewarded stimulus is no longer associated with reward but the previously unrewarded stimulus is now associated with reward. How rapidly an animal learns to reverse their responding to match the change in reward contingencies is a measure of how flexible their behavior is. The stimuli can be from any sensory dimension, smell, audition, somatosensation, and vision. Visual discriminations are used most commonly in primate studies as visual cues are particularly salient for primates. The stimuli are either spatial in nature, i.e., left is rewarded, but right is not, or they involve visual features such as shapes or color. A typical discrimination reversal task is shown in Fig. 1ei where an animal may receive a series of reward contingency reversals.
A selective deficit in cognitive flexibility is shown by intact performance on the original discrimination, ruling out perceptual deficits, and an impairment on the subsequent reversal or series of reversals; a deficit seen following damage to the orbitofrontal cortex and ventromedial striatum. A more fine-grained analysis is then required to determine the nature of the underlying reversal deficit. Some useful information can be gleaned from a careful analysis of the errors that are made while performing the reversal. The pattern of errors on a reversal task is very distinctive. First monkeys tend to respond to the previously rewarded stimulus, almost exclusively, known as the perseverative stage. Then, they respond randomly to both stimuli (chance stage) and finally begin to respond more to the previously unrewarded, but now rewarded stimulus (learning stage). By analyzing the numbers and proportions of errors made at these three stages some insight can be gained as to whether animals have problems (1) disengaging their attention and inhibiting their responding to the previously rewarded stimulus (► perseveration) or alternatively (2) learning to respond to a stimulus that they had previously learnt was not associated with reward (learned avoidance). If it is the former, animals would make more errors in the perseverative stage, but if it is the latter, then they may make more errors in the chance and/or learning stages. It has been highlighted that in the classic discrimination reversal task, there are only two stimuli to choose from, and thus the only error is a response to the previously rewarded stimulus. A better measure of whether the deficit is truly perseverative in nature may be gained by requiring the animal to perform a three-stimulus visual discrimination reversal task as highlighted by JD Jentsch and JR Taylor. In which case, an error would include, not only a response to the previously rewarded stimulus but also a response to the other, previously unrewarded stimulus. Thus, a more generalized impairment in reversal learning may manifest itself as errors made equally to both of the currently, unrewarded stimuli while a per-severative deficit would still be characterized by errors made primarily to the stimulus that had been previously rewarded. This version can rule out perseverative responding as an underlying cause of a reversal deficit.
However, if perseverative responding is seen, the underlying cause of the perseverative deficit is still unclear. It may still be a consequence of the subject avoiding the two, previously unrewarded stimuli, rather than due to a failure to inhibit responding to the previously rewarded stimulus. To differentiate these two possibilities, the following design can be used. At the reversal stage ofthe discrimination task, one of two different versions of the discrimination is given. One version includes the previously rewarded stimulus and a novel stimulus and the monkey has to choose the novel stimulus. The other version includes the previously unrewarded stimulus and a novel stimulus and the monkey has to select the previously unrewarded stimulus (Fig. 1eii). If the perseverative deficit seen in the original, discrimination reversal task is truly perseverative then the monkey should be impaired on the former but not the latter. In contrast, if the deficit is due to the animal actively avoiding the previously unrewarded stimulus, then they should be impaired on the latter and not the former (Clarke et al. 2006).
Object Retrieval test. Another commonly used test of cognitive flexibility that has proven useful in psycho-pharmacological studies is a test of object retrieval in which monkeys have to make a detour reach around the sides of a clear Perspex box in order to retrieve the food reward inside. Performance on this test is dependent upon the orbitofrontal cortex and the caudate nucleus (and also large lesions of dorsolateral prefrontal cortex including Walker's areas 9,46 and 8). It not only investigates the ability of monkeys to inhibit a prepotent response tendency to reach directly for the food reward but also their ability to switch their responding between the left and right sides of the box, as on some trials the opening is on the left and other trials, on the right (Jentsch et al. 1999).
Attentional set (rule) -shifting tasks. The psychological and neural mechanisms underlying flexible responding to changes in stimulus-reward contingencies are not the same as those required for the flexible use of rules to guide responding. One of the most commonly used tests to study rule switching in humans is the Wisconsin Card Sorting Test (WCST), which requires subjects to sort a pack of cards according to one particular perceptual dimension, e.g., shape, and then to switch to sorting them according to another dimension, e.g., number. A number of different versions of this test have been adapted for use in nonhuman primates, the different versions focussing on different aspects of the original task. In one such version (Roberts et al. 1988), monkeys are presented with a series of visual discriminations composed of bidi-mensional stimuli using abstract dimensions of "shape" and "line" (Fig. 1f). The monkey has to learn that one particular perceptual dimension is relevant to the task and that an exemplar from that dimension is associated with food reward, i.e., the blue boat, regardless of the white line superimposed over it. Animals then perform a series ofsuch discriminations in which the same dimension remains relevant throughout but each discrimination is composed of novel bidimensional stimuli in which one exemplar from the relevant dimension is associated with food each time (► intradimensional shift, IDS). In the critical set-shifting test (or ► extradimensional shift, EDS) a discrimination is presented in which the previously relevant dimension is no longer relevant and the subject has to learn, through trial and error, that an exemplar from the previously irrelevant dimension is now rewarded. This requires monkeys to shift their ► attentional set from one dimension to another and is similar to the shift of category in the WCST. Impairment at the EDS stage is dependent upon the lateral PFC. The advantage of this test is that it separates out some of the component parts of the WCST, including the ability to develop an attentional set, to apply the rule across different discriminations and then to switch from using one rule to using another. A distractor probe test, in which the exemplars from the irrelevant dimension of a well-learned discrimination are replaced for one session only with novel irrelevant exemplars, can be used to investigate how much an animal is distracted by the irrelevant dimension. This particular task has a major emphasis on learning. In contrast, other primate rule switching tests are more akin to the WCST and require monkeys to learn to select, from a set of three stimuli, the stimulus that matches the sample stimulus according to a particular perceptual dimension (Man-souri et al. 2006). In these tasks, the monkeys receive extensive training on the matching to sample rules before they are able to perform the task successfully. Both these and the previously described discrimination tests require subjects to use feedback in the form of reward to guide their responding at the time of the shift. In contrast, another set of task-switching paradigms focus on the mechanisms underlying task switching per se, and in these cases, cues signal which particular rule is in operation at any one time (Stoet and Synder 2008).
Was this article helpful?