Monkey prefrontal neurons during Sternberg task performance: full contents of working memory or most recent item?

To explore the brain mechanisms underlying multi-item working memory, we monitored the activity of neurons in the dorsolateral prefrontal cortex while macaque monkeys performed spatial and chromatic versions of a Sternberg working-memory task. Each trial required holding three sequentially presented samples in working memory so as to identify a subsequent probe matching one of them. The monkeys were able to recall all three samples at levels well above chance, exhibiting modest load and recency effects. Prefrontal neurons signaled the identity of each sample during the delay period immediately following its presentation. However, as each new sample was presented, the representation of antecedent samples became weak and shifted to an anomalous code. A linear classifier operating on the basis of population activity during the final delay period was able to perform at approximately the level of the monkeys on trials requiring recall of the third sample but showed a falloff in performance on trials requiring recall of the first or second sample much steeper than observed in the monkeys. We conclude that delay-period activity in the prefrontal cortex robustly represented only the most recent item. The monkeys apparently based performance of this classic working-memory task on some storage mechanism in addition to the prefrontal delay-period firing rate. Possibilities include delay-period activity in areas outside the prefrontal cortex and changes within the prefrontal cortex not manifest at the level of the firing rate.NEW & NOTEWORTHY It has long been thought that items held in working memory are encoded by delay-period activity in the dorsolateral prefrontal cortex. Here we describe evidence contrary to that view. In monkeys performing a serial multi-item working memory task, dorsolateral prefrontal neurons encode almost exclusively the identity of the sample presented most recently. Information about earlier samples must be encoded outside the prefrontal cortex or represented within the prefrontal cortex in a cryptic code.

Stimulus-response incompatibility activates cortex proximate to three eye fields.

We used functional magnetic resonance imaging (fMRI) to investigate cortical activation during the performance of three oculomotor tasks that impose increasing levels of cognitive demand. (1) In a visually guided saccade (VGS) task, subjects made saccades to flashed targets. (2) In a compatible task, subjects made leftward and rightward saccades in response to foveal presentation of the uppercase words "LEFT" or "RIGHT." (3) In a mixed task, subjects made rightward saccades in response to the lowercase word "left" and leftward saccades in response to the lowercase word "right" on incompatible trials (60%). The remaining 40% of trials required compatible responses to uppercase words. The VGS and compatible tasks, when compared to fixation, activated the three cortical eye fields: the supplementary eye field (SEF), the frontal eye field (FEF), and the parietal eye field (PEF). The mixed task, when compared to the compatible task, activated three additional cortical regions proximate to the three eye fields: (1) rostral to the SEF in medial frontal cortex; (2) rostral to the FEF in dorsolateral prefrontal cortex (DLPFC); (3) rostral and lateral to the PEF in posterior parietal cortex. These areas may contribute to the suppression of prepotent responses and in holding novel visuomotor associations in working memory.

Object-based vision and attention in primates.

In forming a representation of a visible object, the brain must analyze the visual scene pre-attentively, select an object through active attention, and form representations of the multiple attributes of the selected object. During the past two years, progress has been made in understanding the neural underpinnings of these processes by means of single-neuron recording in monkeys.

Neuronal activity in macaque supplementary eye field during planning of saccades in response to pattern and spatial cues.

The aim of this study was to determine whether neuronal activity in the macaque supplementary eye field (SEF) is influenced by the rule used for saccadic target selection. Two monkeys were trained to perform a variant of the memory-guided saccade task in which any of four visible dots (rightward, upward, leftward, and downward) could be the target. On each trial, the cue identifying the target was either a spot flashed in superimposition on the target (spatial condition) or a foveally presented digitized image associated with the target (pattern condition). Trials conforming to the two conditions were interleaved randomly. On recording from 439 SEF neurons, we found that two aspects of neuronal activity were influenced by the nature of the cue. 1) Activity reflecting the direction of the impending response developed more rapidly following spatial than following pattern cues. 2) Activity throughout the delay period tended to be higher following pattern than following spatial cues. We consider these findings in relation to the possible involvement of the SEF in processes underlying attention, arousal, response-selection, and motor preparation.

Macaque supplementary eye field neurons encode object-centered locations relative to both continuous and discontinuous objects.

Many neurons in the supplementary eye field (SEF) of the macaque monkey fire at different rates before eye movements to the right or the left end of a horizontal bar regardless of the bar's location in the visual field. We refer to such neurons as carrying object-centered directional signals. The aim of the present study was to throw light on the nature of object-centered direction selectivity by determining whether it depends on the reference image's physical continuity. To address this issue, we recorded from 143 neurons in two monkeys. All of these neurons were located in a region coincident with the SEF as mapped out in previous electrical stimulation studies and many exhibited task-related activity in a standard saccade task. In each neuron, we compared neuronal activity across trials in which the monkey made eye movements to the right or left end of a reference image. On interleaved trials, the reference image might be either a horizontal bar or a pair of discrete dots in a horizontal array. The dominant effect revealed by this experiment was that neurons selectively active before eye movements to the right (or left) end of a bar were also selectively active before eye movements to the right (or left) dot in a horizontal array. An additional minor effect, present in around a quarter of the sample, took the form of a difference in firing rate between bar and dot trials, with the greater level of activity most commonly associated with dot trials. These phenomena could not be accounted for by minor intertrial differences in the physical directions of eye movements. In summary, SEF neurons carry object-centered signals and carry these signals regardless of whether the reference image is physically continuous or disjunct.

Parental care and clutch sizes in North and South American birds.

The evolutionary causes of small clutch sizes in tropical and Southern Hemisphere regions are poorly understood. Alexander Skutch proposed 50 years ago that higher nest predation in the south constrains the rate at which parent birds can deliver food to young and thereby constrains clutch size by limiting the number of young that parents can feed. This hypothesis for explaining differences in clutch size and parental behaviors between latitudes has remained untested. Here, a detailed study of bird species in Arizona and Argentina shows that Skutch's hypothesis explains clutch size variation within North and South America. However, neither Skutch's hypothesis nor two major alternatives explain differences between latitudes.

Mirror-image confusion in single neurons of the macaque inferotemporal cortex.

Humans and animals confuse lateral mirror images, such as the letters "b" and "d," more often than vertical mirror images, such as the letters "b" and "p." Experiments were performed to find a neural correlate of this phenomenon. Visually responsive pattern-selective neurons in the inferotemporal cortex of macaque monkeys responded more similarly to members of a lateral mirror-image pair than to members of a vertical mirror-image pair. The phenomenon developed within 20 milliseconds of the onset of the visual response and persisted to its end. It occurred during presentation of stimuli both at the fovea and in the periphery.

Macaque SEF neurons encode object-centered directions of eye movements regardless of the visual attributes of instructional cues.

Macaque SEF neurons encode object-centered directions of eye movements regardless of the visual attributes of instructional cues. Neurons in the supplementary eye field (SEF) of the macaque monkey exhibit object-centered direction selectivity in the context of a task in which a spot flashed on the right or left end of a sample bar instructs a monkey to make an eye movement to the right or left end of a target bar. To determine whether SEF neurons are selective for the location of the cue, as defined relative to the sample bar, or, alternatively, for the location of the target, as defined relative to the target bar, we carried out recording while monkeys performed a new task. In this task, the color of a cue-spot instructed the monkey to which end of the target bar an eye movement should be made (blue for the left end and yellow for the right end). Object-centered direction selectivity persisted under this condition, indicating that neurons are selective for the location of the target relative to the target bar. However, object-centered signals developed at a longer latency (by approximately 200 ms) when the instruction was conveyed by color than when it was conveyed by the location of a spot on a sample bar.

Representation of object-centered space in the primate frontal lobe.

Object-centered spatial awareness--awareness of locations of parts relative to a an object--plays an important role in perception and action. Indirect evidence from psychological and neuropsychological studies has indicated that this form of spatial awareness may be served by a cortical system in which neurons encode specific object-centered locations. We set out to obtain direct evidence for object-centered spatial selectivity by recording from single neurons in the frontal cortex of monkeys trained to make eye movements to particular locations on reference objects. We found that neurons in the supplementary eye field (SEF) fire differentially as a function of the location on an object to which an eye movement is directed.

Single neurons in posterior cingulate cortex of behaving macaque: eye movement signals.

1. Posterior cingulate cortex, although widely regarded as a part of the limbic system, is connected most strongly to parietal and frontal areas with sensory, motor, and cognitive functions. To gain insight into the functional nature of posterior cingulate cortex, we have recorded from its neurons in monkeys performing oculomotor tasks known to activate parietal and frontal neurons. We have found that posterior cingulate neurons fire during periods of ocular fixation at a rate determined by the angle of gaze and by the size and direction of the preceding eye movement. 2. The activity of 530 posterior cingulate neurons was monitored while rhesus macaque monkeys made visually guided eye movements to spots projected on a tangent screen. 3. In 150/530 neurons, a statistically significant shift in the rate of discharge occurred around the time of onset of saccadic eye movements. The preponderant form of response was an increase in activity (142/150 neurons). 4. In 142 neurons exhibiting significant excitation after saccades in at least one direction, the level of discharge was analyzed as a function of time relative to onset of the saccade. Across the neuronal population as a whole, activity increased sharply at the moment of onset of the saccade, rising to a maximum after 200 ms and then declining slowly. The net level of discharge remained well above presaccadic baseline even after > 1 s of postsaccadic fixation. 5. In 63 neurons, the postsaccadic rate of discharge was analyzed relative to the angle of the eye in the orbit by monitoring neuronal activity while the monkey executed saccades of uniform direction and amplitude to four targets spaced at 16-deg intervals along a line. The postsaccadic firing level was significantly dependent on orbital angle in 44/63 neurons. 6. In 45 neurons, the postsaccadic rate of discharge was analyzed relative to saccade direction by monitoring neuronal activity while the monkey executed 16-deg saccades to a constant target from diametrically opposed starting points. The postsaccadic level of activity was significantly dependent on saccade direction in 20/ 45 neurons. 7. In 58 neurons, the postsaccadic rate of discharge was analyzed relative to saccade amplitude by monitoring neuronal activity while the monkey executed saccades, which varied in amplitude (4, 8, 16, and 32 deg) but which were constant in direction and brought the eye to bear on a constant endpoint. The postsaccadic level of activity was significantly dependent on saccade amplitude in 24/58 neurons. In all neurons exhibiting significant amplitude-dependence, stronger firing accompanied larger saccades. 8. The activity of 10 neurons was monitored during smooth pursuit eye movements (20 deg/s upward, downward, leftward, and rightward). The level of firing varied as a function of both the position of the eye (9 neurons) and the velocity of the eye (6 neurons). 9. We conclude that posterior cingulate neurons monitor eye movements and eye position. It is unlikely that they participate in the generation of eye movements because their shifts of discharge follow the onset of the movements. Eye-movement-related signals in posterior cingulate cortex may reflect the participation of this area in assigning spatial coordinates to retinal images.

Brain representation of object-centered space.

Object-centered spatial awareness underlies many important cognitive functions, including reading, which requires registering the locations of letters relative to a word, and pattern recognition, which requires registering the locations of features relative to a whole pattern. Recent studies have elucidated the nature of the brain mechanisms underlying this form of spatial awareness by showing the attention tends to focus on objects rather than on regions of space: by demonstrating that each hemisphere contributes selectively to awareness of the opposite half of object space, and by revealing that neurons in some cortical areas are selective for particular locations in object space. These results are concordant with the general idea that imagining or attending to an object is accompanied by projecting its image onto a neural map of object-centered space. An important aim for future studies will be to test and extend this 'object map' hypothesis.

Object-centered direction selectivity in the macaque supplementary eye field.

Object-centered spatial awareness--awareness of the location, relative to an object, of its parts--plays an important role in many aspects of perception, imagination, and action. One possible basis for this capability is the existence in the brain of neurons with sensory receptive fields or motor action fields that are defined relative to an object-centered frame. In experiments described here, neuronal activity was monitored in the supplementary eye field of macaque monkeys making eye movements to the right or left end of a horizontal bar. Neurons were found to fire differentially as a function of the end of the bar to which an eye movement was made. This is direct evidence for the existence of neurons sensitive to the object-centered direction of movements.

The role of striate cortex in visual function of the cat.

We examined the contribution of area 17 to visual function in two cats whose fixation was monitored by means of scleral search coils. Ibotenic acid lesions were made within the physiologically identified representation of the lower left visual field of area 17. In a detection task in which the cats simply indicated the presence or absence of a vertical grating, contrast sensitivity loss was greatest at middle spatial frequencies with no loss in spatial resolution. However, when cats were required to discriminate between vertical and horizontal gratings, sensitivity loss was profound at both middle and high spatial frequencies with an octave loss of spatial resolution. This greater loss was not due to disrupted orientation discrimination since sensitivity to the orientation of coarse gratings was unaffected in the lesioned hemifield. We also found deficits in the ability to discriminate the direction of grating motion, but only at higher spatial and lower temporal frequencies. The role of area 17 in perceiving the global motion of complex patterns was also studied with high contrast, dynamic random dots drifting at high speeds. Paradoxically, area 17 lesion improved the perception of global motion. This improvement was eliminated by spatially filtering the dot patterns to remove high spatial frequencies, suggesting that the lesion has enhanced performance by interfering with masking by high spatial frequencies. Our results demonstrate that the performance of traditional detection tasks may be insensitive to the effects of area 17 lesions. Discrimination tasks, on the other hand, revealed that area 17 neurons play a major role in the perception of higher spatial frequency stimuli as long as they move or flicker at low rates, but contribute little to these functions when the stimuli are coarse and move at high speeds.

Functional heterogeneity in cingulate cortex: the anterior executive and posterior evaluative regions.

The cingulate gyrus is a major part of the "anatomical limbic system" and, according to classic accounts, is involved in emotion. This view is oversimplified in light of recent clinical and experimental findings that cingulate cortex participates not only in emotion but also in sensory, motor, and cognitive processes. Anterior cingulate cortex, consisting of areas 25 and 24, has been implicated in visceromotor, skeletomotor, and endocrine outflow. These processes include responses to painful stimuli, maternal behavior, vocalization, and attention to action. Since all of these activities have an affective component, it is likely that connections with the amygdala are critical for them. In contrast, posterior cingulate cortex, consisting of areas 29, 30, 23, and 31, contains neurons that monitor eye movements and respond to sensory stimuli. Ablation studies suggest that this region is involved in spatial orientation and memory. It is likely that connections between posterior cingulate and parahippocampal cortices contribute to these processes. We conclude that there is a fundamental dichotomy between the functions of anterior and posterior cingulate cortices. The anterior cortex subserves primarily executive functions related to the emotional control of visceral, skeletal, and endocrine outflow. The posterior cortex subserves evaluative functions such as monitoring sensory events and the organism's own behavior in the service of spatial orientation and memory.

Posterior cingulate cortex: sensory and oculomotor properties of single neurons in behaving cat.

The posterior cingulate cortex of the cat is strongly linked to cortical areas with sensory and oculomotor functions. We have now recorded from feline posterior cingulate neurons in order to determine whether they are active in conjunction with sensory events and eye movements. The results described here are based on monitoring the electrical activity of 195 single neurons in the posterior cingulate cortex of three cats equipped with surgically implanted scleral search coils and trained to fixate visual targets. Posterior cingulate neurons carry tonic orbital position signals and are phasically active in conjunction with saccadic eye movements. Activity related to eye movements and gaze is attenuated but not abolished by the elimination of visual feedback. Posterior cingulate neurons also are responsive to visual, auditory, and somatosensory stimulation. Systematic testing with visual stimuli revealed that responses are sharply reduced due to refractoriness at rates of stimulation greater than a few per second. These results conform to the theory that posterior cingulate cortex is involved in processes underlying visuospatial cognition.

Topographic organization of cortical and subcortical projections to posterior cingulate cortex in the cat: evidence for somatic, ocular, and complex subregions.

The posterior cingulate area (CGp) of the cat consists of cortex on the exposed cingulate gyrus and in the adjacent ventral bank of the splenial sulcus. We have placed deposits of distinguishable fluorescent tracers at multiple restricted sites in CGp and have analyzed the distribution throughout the forebrain of neurons labeled by retrograde transport. Cortical projections to CGp arise (in approximately descending order of strength) from anterior cingulate cortex; prefrontal cortex and premotor areas including the frontal eye fields; visual areas including especially areas 7 and 20b; parahippocampal areas; insular cortex; somesthetic areas; and auditory areas. Corticocortical pathways are organized topographically with respect to the posterior-anterior axis in CGp. Projections from prefrontal cortex and other areas with complex (as opposed to sensory, motor, or limbic) functions are concentrated posteriorly; projections from visual and oculomotor areas are concentrated at an intermediate level; and projections from areas with somesthetic and somatomotor functions are concentrated anteriorly. Thalamic projections to CGp arise from the anterior nuclei (AD, AV, and AM), from restricted portions of the ventral complex (VAd, VAm, and VMP), from discrete sectors of the lateral complex (LD, LPs, and LPm), from the rostral crescent of intralaminar nuclei (CM, PC, and CL), and from the reuniens nucleus. Projections from AM, VAd, LD, and LPs are spatially ordered in the sense that more ventral thalamic neurons project to more anterior cortical sites. Projections from AV and AD are stronger at more posterior cortical sites but do not show other signs of topographic ordering. Projections from LPm, CM, PC, CL, and RE are diffuse. We conclude (1) that cortical afferents of CGp derive predominantly from neocortical areas including those with well established sensory and motor functions; (2) that limbic projections to CGp originate primarily in structures, including the hippocampus, which are associated with memory, as opposed to structures, including the amygdala, which are associated with emotional and instinctual behavior; and (3) that CGp contains subregions in which complex, ocular, or somatic afferents predominate.

Cortical areas in the medial frontal lobe of the cat delineated by quantitative analysis of thalamic afferents.

The aim of these experiments was to establish the number and location of connectionally distinct areas in the medial frontal lobe of the cat. Thirty deposits of distinguishable retrograde tracers were placed at restricted sites spanning the medial frontal lobe in 16 cats. Following each deposit, the number of retrogradely labeled neurons in each of 17 thalamic nuclei was determined. Variations of the thalamic labeling pattern dependent on the location of the cortical tracer deposit were then analyzed by a quantitative procedure. The results indicate that the medial frontal lobe contains three fundamental divisions: the anterior cingulate area, medial area 6, and the medial prefrontal district. The anterior cingulate area derives its strongest thalamic input from the anteriomedial nucleus. Medial area 6 is the target of afferents originating in a dorsolateral sector of the mediodorsal nucleus and in the ventroanterior nucleus. Medial prefrontal cortex is heavily innervated by pathways originating in the core of the mediodorsal nucleus and in the principal ventromedial nucleus. Within each major district, thalamic connectional patterns exhibit graded regional variation, with the result that, whereas the connections of the district are not uniform, it is difficult to define further discrete subdivisions. We discuss these results in relation to previously proposed schemes for paracellation of the cat's medial frontal lobe and conclude that the infralimbic and prelimbic areas (areas 25 and 32) of previous systems are best understood not as discrete areas but as ventral and intermediate sectors of a continuous medial prefrontal domain.

Visual and auditory association areas of the cat's posterior ectosylvian gyrus: thalamic afferents.

The feline posterior ectosylvian gyrus contains a broad band of association cortex that is bounded anteriorly by tonotopic auditory areas and posteriorly by retinotopic visual areas. To characterize the possible functions of this cortex and to throw light on its pattern of internal divisions, we have carried out an analysis of its thalamic afferents. Deposits of differentiable retrograde tracers were placed at 17 cortical sites in nine cats. The deposit sites spanned the crown of the posterior ectosylvian gyrus and adjacent cortex in the suprasylvian sulcus. We compiled counts of retrogradely labeled neurons in 12 thalamic nuclei delineated by use of Nissl and acetylcholinesterase stains. We then employed a statistical clustering algorithm to identify groups of injections that gave rise to similar patterns of thalamic labeling. The results suggest that the posterior ectosylvian gyrus contains 3 fundamentally different cortical districts that have the form of parallel vertical bands. Very anterior cortex, overlapping previously identified tonotopic auditory areas (AI, P and VP) receives a dense projection from the laminated division of the medial geniculate body (MGl). An intermediate strip, to which we refer as the auditory belt, is innervated by axons from nontonotopic divisions of the medial geniculate body (MGds, MGvl, MGm, and MGd), from the lateral division of the posterior group (Pol), and from the posterior suprageniculate nucleus (SGp). A posterior strip, to which we refer as EPp, receives strong projections from the LM-SG complex (LM-SGa and LMp), and lighter projections from the intralaminar and lateroposterior (LPm and LPl) nuclei. On grounds of thalamic connectivity, EPp is not obviously distinguishable from adjacent retinotopic visual areas (PLLS, DLS, and VLS), and may be regarded as forming, together with these areas, a connectionally homogeneous visual belt.

Visual and auditory association areas of the cat's posterior ectosylvian gyrus: cortical afferents.

In a preceding report, we described patterns of thalamic retrograde labeling following 17 tracer deposits on the cat's posterior ectosylvian gyrus and concluded, on the basis of patterns of thalamic connectivity, that the posterior ectosylvian gyrus is composed of three major divisions: a tonotopic auditory zone located anteriorly, a belt of auditory association cortex occupying the gyral crown, and a visual belt located posteriorly. We describe here patterns of transcortical retrograde labeling obtained from tracer deposits in the three zones so defined. Our results indicate that the tonotopic auditory strip is innervated primarily by axons from low-order auditory areas (AAF, AI, P, VP, and V), that the auditory belt receives its strongest input from nontonotopic auditory fields (AII, temporal cortex, and other parts of the auditory belt), and that projections to the visual belt derive primarily from extrastriate visual areas (ALLS, PLLS, DLS, 19, 20, and 21) and from association areas affiliated with the visual system (insular cortex, posterior cingulate gyrus, area 7p, and frontal cortex). We discuss the results in relation to previous systems for parcellating the posterior ectosylvian gyrus of the cat and consider the possibility that divisions of the feline posterior ectosylvian gyrus correspond directly to areas making up the superior temporal gyrus in primates.

Organization of cortical and subcortical projections to anterior cingulate cortex in the cat.

The aim of the experiments reported here was to identify cortical and subcortical forebrain structures from which anterior cingulate cortex (CGa) receives input in the cat. Deposits of retrograde tracers were placed at nine sites spanning the anterior cingulate area and patterns of retrograde transport were analyzed. Thalamic projections to CGa, in descending order of strength, originate in the anteromedial nucleus, lateroposterior nucleus, ventroanterior nucleus, rostral intralaminar complex, reuniens nucleus, mediodorsal nucleus, and laterodorsal nucleus. Minor and inconsistent ascending pathways arise in the paraventricular, parataenial, parafascicular, and subparafascicular thalamic nuclei. The basolateral nucleus of the amygdala, the hypothalamus, the nucleus of the diagonal band, and the claustrum are additional sources of ascending input. Cortical projections to CGa, in descending order of strength, derive from posterior cingulate cortex, prefrontal cortex, motor cortex (areas 4 and 6), parahippocampal cortex (entorhinal, perirhinal, postsubicular, parasubicular, and subicular areas), insular cortex, somesthetic cortex (areas 5 and SIV), and visual cortex (areas 7p, 20b, AMLS, PS and EPp). In general, the limbic, sensory, and motor afferents of CGa are weak. The dominant sources of input to CGa are other cortical areas with high-order functions. This finding calls into question the traditional characterization of cingulate cortex as a bridge between neocortical association areas and the limbic system.