Taste and olfactory status in a gourmand with a right amygdala lesion.

In a patient with a lesion of the right amygdala and temporal pole who had the characteristics of the gourmand syndrome, sensory and hedonic testing was performed to examine the processing of taste, olfactory, and some emotional stimuli. The gourmand syndrome describes a preoccupation with food and a preference for fine eating and is associated with right anterior lesions. It was found that the taste thresholds for sweet, salt, bitter, and sour were normal; that the patient did not dislike the taste of salt (NaCl) at low and moderate concentrations as much as age-matched controls; that this also occurred for monosodium glutamate (MSG); that there were some olfactory differences from normal controls; and that there was a marked reduction in the ability to detect face expressions of disgust.

Taste, olfactory and food texture reward processing in the brain and obesity.

Complementary neuronal recordings and functional neuroimaging in humans, show that the primary taste cortex in the anterior insula provides separate and combined representations of the taste, temperature and texture (including fat texture) of food in the mouth independently of hunger and thus of reward value and pleasantness. One synapse on, in the orbitofrontal cortex (OFC), these sensory inputs are for some neurons combined by learning with olfactory and visual inputs, and these neurons encode food reward in that they only respond to food when hungry, and in that activations correlate with subjective pleasantness. Cognitive factors, including word-level descriptions, and attention, modulate the representation of the reward value of food in the OFC. Further, there are individual differences in the representation of the reward value of food in the OFC. It is argued that overeating and obesity are related in many cases to an increased reward value of the sensory inputs produced by foods, and their modulation by cognition and attention, which overrides existing satiety signals. It is proposed that control of all rather than one or several of these factors that influence food reward and eating may be important in the prevention and treatment of overeating and obesity.

Continuous transformation learning of translation invariant representations.

We show that spatial continuity can enable a network to learn translation invariant representations of objects by self-organization in a hierarchical model of cortical processing in the ventral visual system. During 'continuous transformation learning', the active synapses from each overlapping transform are associatively modified onto the set of postsynaptic neurons. Because other transforms of the same object overlap with previously learned exemplars, a common set of postsynaptic neurons is activated by the new transforms, and learning of the new active inputs onto the same postsynaptic neurons is facilitated. We show that the transforms must be close for this to occur; that the temporal order of presentation of each transformed image during training is not crucial for learning to occur; that relatively large numbers of transforms can be learned; and that such continuous transformation learning can be usefully combined with temporal trace training.

Functions of the orbitofrontal and pregenual cingulate cortex in taste, olfaction, appetite and emotion.

Complementary neurophysiological recordings in macaques and functional neuroimaging in humans show that the primary taste cortex in the rostral insula and adjoining frontal operculum provides separate and combined representations of the taste, temperature, and texture (including viscosity and fat texture) of food in the mouth independently of hunger and thus of reward value and pleasantness. One synapse on, in the orbitofrontal cortex, these sensory inputs are for some neurons combined by learning with olfactory and visual inputs. Different neurons respond to different combinations, providing a rich representation of the sensory properties of food. The representation of taste and other food-related stimuli in the orbitofrontal cortex of macaques is found from its lateral border throughout area 13 to within 7 mm of the midline, and in humans the representation of food-related and other pleasant stimuli is found particularly in the medial orbitofrontal cortex. In the orbitofrontal cortex, feeding to satiety with one food decreases the responses of these neurons to that food, but not to other foods, showing that sensory-specific satiety is computed in the primate (including human) orbitofrontal cortex. Consistently, activation of parts of the human orbitofrontal cortex correlates with subjective ratings of the pleasantness of the taste and smell of food. Cognitive factors, such as a word label presented with an odour, influence the pleasantness of the odour, and the activation produced by the odour in the orbitofrontal cortex. Food intake is thus controlled by building a multimodal representation of the sensory properties of food in the orbitofrontal cortex, and gating this representation by satiety signals to produce a representation of the pleasantness or reward value of food which drives food intake. A neuronal representation of taste is also found in the pregenual cingulate cortex, which receives inputs from the orbitofrontal cortex, and in humans many pleasant stimuli activate the pregenual cingulate cortex, pointing towards this as an important area in motivation and emotion.

Learning transform invariant object recognition in the visual system with multiple stimuli present during training.

Over successive stages, the visual system develops neurons that respond with view, size and position invariance to objects or faces. A number of computational models have been developed to explain how transform-invariant cells could develop in the visual system. However, a major limitation of computer modelling studies to date has been that the visual stimuli are typically presented one at a time to the network during training. In this paper, we investigate how vision models may self-organize when multiple stimuli are presented together within each visual image during training. We show that as the number of independent stimuli grows large enough, standard competitive neural networks can suddenly switch from learning representations of the multi-stimulus input patterns to representing the individual stimuli. Furthermore, the competitive networks can learn transform (e.g. position or view) invariant representations of the individual stimuli if the network is presented with input patterns containing multiple transforming stimuli during training. Finally, we extend these results to a multi-layer hierarchical network model (VisNet) of the ventral visual system. The network is trained on input images containing multiple rotating 3D objects. We show that the network is able to develop view-invariant representations of the individual objects.

Invariant object recognition with trace learning and multiple stimuli present during training.

Over successive stages, the ventral visual system develops neurons that respond with view, size and position invariance to objects including faces. A major challenge is to explain how invariant representations of individual objects could develop given visual input from environments containing multiple objects. Here we show that the neurons in a 1-layer competitive network learn to represent combinations of three objects simultaneously present during training if the number of objects in the training set is low (e.g. 4), to represent combinations of two objects as the number of objects is increased to for e.g. 10, and to represent individual objects as the number of objects in the training set is increased further to for e.g. 20. We next show that translation invariant representations can be formed even when multiple stimuli are always present during training, by including a temporal trace in the learning rule. Finally, we show that these concepts can be extended to a multi-layer hierarchical network model (VisNet) of the ventral visual system. This approach provides a way to understand how a visual system can, by self-organizing competitive learning, form separate invariant representations of each object even when each object is presented in a scene with multiple other objects present, as in natural visual scenes.

Learning movement sequences with a delayed reward signal in a hierarchical model of motor function.

A key problem in reinforcement learning is how an animal is able to learn a sequence of movements when the reward signal only occurs at the end of the sequence. We describe how a hierarchical dynamical model of motor function is able to solve the problem of delayed reward in learning movement sequences using associative (Hebbian) learning. At the lowest level, the motor system encodes simple movements or primitives, while at higher levels the system encodes sequences of primitives. During training, the network is able to learn a high level motor program composed of a specific temporal sequence of motor primitives. The network is able to achieve this despite the fact that the reward signal, which indicates whether or not the desired motor program has been performed correctly, is received only at the end of each trial during learning. Use of a continuous attractor network in the architecture enables the network to generate the motor outputs required to produce the continuous movements necessary to implement the motor sequence.

Learning invariant object recognition in the visual system with continuous transformations.

The cerebral cortex utilizes spatiotemporal continuity in the world to help build invariant representations. In vision, these might be representations of objects. The temporal continuity typical of objects has been used in an associative learning rule with a short-term memory trace to help build invariant object representations. In this paper, we show that spatial continuity can also provide a basis for helping a system to self-organize invariant representations. We introduce a new learning paradigm "continuous transformation learning" which operates by mapping spatially similar input patterns to the same postsynaptic neurons in a competitive learning system. As the inputs move through the space of possible continuous transforms (e.g. translation, rotation, etc.), the active synapses are modified onto the set of postsynaptic neurons. Because other transforms of the same stimulus overlap with previously learned exemplars, a common set of postsynaptic neurons is activated by the new transforms, and learning of the new active inputs onto the same postsynaptic neurons is facilitated. We demonstrate that a hierarchical model of cortical processing in the ventral visual system can be trained with continuous transform learning, and highlight differences in the learning of invariant representations to those achieved by trace learning.

The perirhinal cortex and long-term familiarity memory.

To analyse the functions of the perirhinal cortex, the activity of single neurons in the perirhinal cortex was recorded while macaques performed a delayed matching-to-sample task with up to three intervening stimuli. Some neurons had activity related to working memory, in that they responded more to the sample than to the match image within a trial, as shown previously. However, when a novel set of stimuli was introduced, the neuronal responses were on average only 47% of the magnitude of the responses to the set of very familiar stimuli. Moreover, it was shown in three monkeys that the responses of the perirhinal cortex neurons gradually increased over hundreds of presentations (mean = 400 over 7-13 days) of the new set of (initially novel) stimuli to become as large as those to the already familiar stimuli. Thus perirhinal cortex neurons represent the very long-term familiarity of visual stimuli. Part of the impairment in temporal lobe amnesia may be related to the difficulty of building representations of the degree of familiarity of stimuli. A neural network model of how the perirhinal cortex could implement long-term familiarity memory is proposed using Hebbian associative learning.

The primate amygdala and reinforcement: a dissociation between rule-based and associatively-mediated memory revealed in neuronal activity.

To investigate the role of the primate amygdala in stimulus-reinforcement association learning, the activity of single amygdala neurons was recorded in macaques during two memory tasks. In a visual discrimination task, a population of neurons (17/659) was analyzed which responded differentially to a visual stimulus which always indicated that the primary reinforcer fruit juice could be obtain if the monkey licked, and a different visual stimulus that indicated that the primary reinforcer aversive saline would be obtained if the monkey licked. Most (16/17) of these neurons responded more to the reward-related than the aversive visual stimulus. In a recognition memory task, the majority (12/14 analyzed) of these neurons responded equally well to the trial unique stimuli when they were shown as novel and the monkey had to not lick in order to avoid saline, and when they were shown a second time as familiar and the monkey used the rule that if he licked, fruit juice would be obtained. The responses of these amygdala neurons thus reflect the direct associations of stimuli with reinforcement, but do not reflect the reward value of the stimuli when this must be assessed based on a rule (in the recognition memory task, that a stimulus will be punished the first time it is shown, and rewarded the second). This finding also shows that these amygdala neurons respond to relatively novel stimuli in the same way as they do to stimuli that have become rewarding by stimulus-reinforcement association learning. This provides a neural basis for relatively novel stimuli to be treated as rewarding, and approached.

The primate amygdala: Neuronal representations of the viscosity, fat texture, temperature, grittiness and taste of foods.

The primate amygdala is implicated in the control of behavioral responses to foods and in stimulus-reinforcement learning, but only its taste representation of oral stimuli has been investigated previously. Of 1416 macaque amygdala neurons recorded, 44 (3.1%) responded to oral stimuli. Of the 44 orally responsive neurons, 17 (39%) represent the viscosity of oral stimuli, tested using carboxymethyl-cellulose in the range 1-10,000 cP. Two neurons (5%) responded to fat in the mouth by encoding its texture (shown by the responses of these neurons to a range of fats, and also to non-fat oils such as silicone oil ((Si(CH(3))(2)O)(n)) and mineral oil (pure hydrocarbon), but no or small responses to the cellulose viscosity series or to the fatty acids linoleic acid and lauric acid). Of the 44 neurons, three (7%) responded to gritty texture (produced by microspheres suspended in cellulose). Eighteen neurons (41%) responded to the temperature of liquid in the mouth. Some amygdala neurons responded to capsaicin, and some to fatty acids (but not to fats in the mouth). Some amygdala neurons respond to taste, texture and temperature unimodally, but others combine these inputs. These results provide fundamental evidence about the information channels used to represent the texture and flavor of food in a part of the brain important in appetitive responses to food and in learning associations to reinforcing oral stimuli, and are relevant to understanding the physiological and pathophysiological processes related to food intake, food selection, and the effects of variety of food texture in combination with taste and other inputs on food intake.

Self-organizing continuous attractor network models of hippocampal spatial view cells.

Single neuron recording studies have demonstrated the existence of hippocampal spatial view neurons which encode information about the spatial location at which a primate is looking in the environment. These neurons are able to maintain their firing even in the absence of visual input. The standard neuronal network approach to model networks with memory that represent continuous spaces is that of continuous attractor networks. It has recently been shown how idiothetic (self-motion) inputs could update the activity packet of neuronal firing for a one-dimensional case (head direction cells), and for a two-dimensional case (place cells which represent the place where a rat is located). In this paper, we describe three models of primate hippocampal spatial view cells, which not only maintain their spatial firing in the absence of visual input, but can also be updated in the dark by idiothetic input. The three models presented in this paper represent different ways in which a continuous attractor network could integrate a number of different kinds of velocity signal (e.g., head rotation and eye movement) simultaneously. The first two models use velocity information from head angular velocity and from eye velocity cells, and make use of a continuous attractor network to integrate this information. A fundamental feature of the first two models is their use of a 'memory trace' learning rule which incorporates a form of temporal average of recent cell activity. Rules of this type are able to build associations between different patterns of neural activities that tend to occur in temporal proximity, and are incorporated in the model to enable the recent change in the continuous attractor to be associated with the contemporaneous idiothetic input. The third model uses positional information from head direction cells and eye position cells to update the representation of where the agent is looking in the dark. In this case the integration of idiothetic velocity signals is performed in the earlier layer of head direction cells.

Orbitofrontal cortex: neuronal representation of oral temperature and capsaicin in addition to taste and texture.

The primate orbitofrontal cortex is a site of convergence of information from primary taste, olfactory and somatosensory cortical areas. We describe the discovery of a population of single neurons in the macaque orbitofrontal cortex that responds to the temperature of a liquid in the mouth. The temperature stimuli consisted of water at 10 degrees C, 23 degrees C, 37 degrees C and 42 degrees C. Twenty-six of the 1149 neurons analyzed (2.3%) responded to oral temperature. The tuning profiles of the neurons to temperature showed that some of the neurons had graded responses to increasing temperature (27%), others responded to cold (10 degrees C) stimuli (27%), and others were tuned to temperature (46%). The neuronal responses were also measured to taste stimuli, viscosity stimuli (carboxymethyl-cellulose in the range 1-10,000 cP), and capsaicin (10 microM). Of 70 neurons with responses to any of these stimuli, 7.1% were unimodal temperature; 11.3% were temperature and taste-sensitive; 7.1% were temperature and viscosity-sensitive; and 11.3% were temperature, taste and viscosity sensitive. Capsaicin activated 15.7% of the population of responsive neurons tested. These results provide the first evidence of how the temperature of what is in the mouth is represented at the neuronal level in the orbitofrontal cortex and the first evidence for any primate cortical area that in some cases this information converges onto single neurons with inputs produced by other sensory properties of food, including taste and texture. The results provide a basis for understanding how particular combinations of oral temperature, taste, and texture can influence the palatability of foods.

Reward-related reversal learning after surgical excisions in orbito-frontal or dorsolateral prefrontal cortex in humans.

Neurophysiological studies in primates and neuroimaging studies in humans suggest that the orbito-frontal cortex is involved in representing the reward value of stimuli and in the rapid learning and relearning of associations between visual stimuli and rewarding or punishing outcomes. In the present study, we tested patients with circumscribed surgical lesions in different regions of the frontal lobe on a new visual discrimination reversal test, which, in an fMRI study (O'Doherty, Kringelbach, Rolls, Hornak, & Andrews, 2001), produced bilateral orbito-frontal cortex activation in normal subjects. In this task, touching one of two simultaneously presented patterns produced reward or loss of imaginary money delivered on a probabilistic basis to minimize the usefulness of verbal strategies. A number of types of feedback were present on the screen. The main result was that the group of patients with bilateral orbito-frontal cortex lesions were severely impaired at the reversal task, in that they accumulated less money. These patients often failed to switch their choice of stimulus after a large loss and often did switch their choice although they had just received a reward. The investigation showed that bilateral lesions were required for this deficit, since patients with unilateral orbito-frontal cortex (or medial prefrontal cortex) lesions were not impaired in the probabilistic reversal task. The task ruled out a simple motor disinhibition as an explanation of the deficit in the bilateral orbito-frontal cortex patients, in that the patients were required to choose one of two stimuli on each trial. A comparison group of patients with dorsolateral prefrontal cortex lesions was in some cases able to do the task, and in other cases, was impaired. Posttest debriefing showed that all the dorsolateral prefrontal patients who were impaired at the task had failed to pay attention to the crucial feedback provided on the screen after each trial about the amount won or lost on each trial. In contrast, all dorsolateral patients who paid attention to this crucial feedback performed normally on the reversal task. Further, it was confirmed that the bilateral orbito-frontal cortex patients had also paid attention to this crucial feedback, but in contrast had still performed poorly at the task. The results thus show that the orbital prefrontal cortex is required bilaterally for monitoring changes in the reward value of stimuli and using this to guide behavior in the task; whereas the dorsolateral prefrontal cortex, if it produces deficits in the task, does so for reasons related to executive functions, such as the control of attention. Thus, the ability to determine which information is relevant when making a choice of pattern can be disrupted by a dorsolateral lesion on either side, whereas the ability to use this information to guide behavior is not disrupted by a unilateral lesion in either the left or the right orbito-frontal cortex, but is severely impaired by a bilateral lesion in this region. Because both abilities are important in many of the tasks and decisions that arise in the course of daily life, the present results are relevant to understanding the difficulties faced by patients after surgical excisions in different frontal brain regions.

Impulsivity, time perception, emotion and reinforcement sensitivity in patients with orbitofrontal cortex lesions.

Damage to the orbitofrontal cortex (OFC) in humans has been associated with disinhibited or socially inappropriate behaviour and emotional changes. Some of the changes may be related to difficulty in responding correctly to rewards and punishers, in that these patients have difficulty in learning to correct their choice of a visual stimulus when it is no longer associated with reward. We extend this fundamental approach by investigating the relationship between frontal dysfunction and impulsive behaviour, the behavioural, emotional and personality changes seen in patients with prefrontal cortex damage, and thus in addition illuminate the cognitive and biological processes that are impaired in impulsive people. OFC patients (n = 23) performed more impulsively on both self-report and cognitive/behavioural tests of impulsivity, reported more inappropriate 'frontal' behaviours, and performed worse on a stimulus-reinforcement association reversal task, than non-OFC prefrontal cortex lesion control (n = 20) and normal control (n = 39) participants. Further, OFC patients experienced more subjective anger than non-OFC and normal participants, and less subjective happiness than normals; and had a faster subjective sense of time (overestimated and underproduced time intervals) than normal controls, while non-OFC patients did not differ from normals. Finally, both OFC and non-OFC patients were less open to experience than normal participants. There were no differences between OFC patients, non-OFC lesion patients and normal controls on all other personality traits, most notably extraversion. In a spatial working memory task, the non-OFC group, most of whom had dorsolateral prefrontal cortex lesions, were impaired in that they repeatedly returned to previously chosen empty locations ('within errors'), whereas OFC patients were not impaired on this measure. Thus there is a dissociation between the effects of OFC damage which does not affect this measure of spatial working memory but does affect impulsive and inappropriate behaviour, reversal, personality, time perception and emotion; and dorsolateral prefrontal cortex damage which does affect this measure of spatial working memory, but not impulsive and inappropriate behaviour, reversal, personality, time perception and emotion. The effects of OFC damage on impulsive and related behaviours described here have implications for understanding impulsive behaviour.

Self-organising continuous attractor networks with multiple activity packets, and the representation of space.

'Continuous attractor' neural networks can maintain a localised packet of neuronal activity representing the current state of an agent in a continuous space without external sensory input. In applications such as the representation of head direction or location in the environment, only one packet of activity is needed. For some spatial computations a number of different locations, each with its own features, must be held in memory. We extend previous approaches to continuous attractor networks (in which one packet of activity is maintained active) by showing that a single continuous attractor network can maintain multiple packets of activity simultaneously, if each packet is in a different state space or map. We also show how such a network could by learning self-organise to enable the packets in each space to be moved continuously in that space by idiothetic (motion) inputs. We show how such multi-packet continuous attractor networks could be used to maintain different types of feature (such as form vs colour) simultaneously active in the correct location in a spatial representation. We also show how high-order synapses can improve the performance of these networks, and how the location of a packet could be read by motor networks. The multiple packet continuous attractor networks described here may be used for spatial representations in brain areas such as the parietal cortex and hippocampus.

Activation of the human orbitofrontal cortex to a liquid food stimulus is correlated with its subjective pleasantness.

Single-neuron recording studies in non-human primates indicate that orbitofrontal cortex neurons represent the reward value of the sight, smell and taste of food, and even changes in the relative reward value, but provide no direct evidence on brain activity that is correlated with subjective reports of the pleasantness of food. In this fMRI investigation we report a significant correlation between the activation of a region of the human orbitofrontal cortex and the decrease in subjective pleasantness when a liquid food is eaten to satiety. Moreover, a cluster of voxels in the orbitofrontal cortex showed a decrease in its activation that was specific to the particular liquid food consumed in a meal, providing a neural correlate of sensory-specific satiety to a liquid whole food in humans. This sensory-specific reduction in activation of the orbitofrontal cortex correlating with subjective pleasantness is consistent with an important role for the orbitofrontal cortex in human emotion and motivation, and associated subjective states.

Representation of umami taste in the human brain.

Umami taste stimuli, of which an exemplar is monosodium glutamate (MSG) and which capture what is described as the taste of protein, were shown using functional MRI (fMRI) to activate similar cortical regions of the human taste system to those activated by a prototypical taste stimulus, glucose. These taste regions included the insular/opercular cortex and the caudolateral orbitofrontal cortex. A part of the rostral anterior cingulate cortex (ACC) was also activated. When the nucleotide 0.005 M inosine 5'-monophosphate (IMP) was added to MSG (0.05 M), the blood oxygenation-level dependent (BOLD) signal in an anterior part of the orbitofrontal cortex showed supralinear additivity; this may reflect the subjective enhancement of umami taste that has been described when IMP is added to MSG. These results extend to humans previous studies in macaques showing that single neurons in these taste cortical areas can be tuned to umami stimuli.

Changes in emotion after circumscribed surgical lesions of the orbitofrontal and cingulate cortices.

To analyse the functions of different parts of the prefrontal cortex in emotion, patients with different prefrontal surgical excisions were compared on four measures of emotion: voice and face emotional expression identification, social behaviour, and the subjective experience of emotion. Some patients with bilateral lesions of the orbitofrontal cortex (OFC) had deficits in voice and face expression identification, and the group had impairments in social behaviour and significant changes in their subjective emotional state. Some patients with unilateral damage restricted to the OFC also had deficits in voice expression identification, and the group did not have significant changes in social behaviour or in their subjective emotional state. Patients with unilateral lesions of the antero-ventral part of the anterior cingulate cortex (ACC) and/or medial Brodmann area (BA) 9 were, in some cases, impaired on voice and face expression identification, had some change in social behaviour, and had significant changes in their subjective emotional state. Patients with unilateral lesions of the OFC and of the ACC and/or medial BA 9 were, in some cases, impaired on voice and face expression identification, had some changes in social behaviour, and had significant changes in their subjective emotional state. Patients with dorsolateral prefrontal cortex lesions or with medial lesions outside ACC and medial BA 9 areas (dorsolateral/other medial group) were unimpaired on any of these measures of emotion. In all cases in which voice expression identification was impaired, there were no deficits in control tests of the discrimination of unfamiliar voices and the recognition of environmental sounds. Thus bilateral or unilateral lesions circumscribed surgically within the OFC can impair emotional voice and/or face expression identification, but significant changes in social behaviour and in subjective emotional state are related to bilateral lesions. Importantly, unilateral lesions of the ACC (including some of medial BA 9) can produce voice and/or face expression identification deficits, and marked changes in subjective emotional state. These findings with surgically circumscribed lesions show that within the prefrontal cortex, both the OFC and the ACC/medial BA 9 region are involved in a number of aspects of emotion in humans including emotion identification, social behaviour and subjective emotional state, and that the dorsolateral prefrontal areas are not involved in emotion in these ways.