Autonomic arousal in an appetitive context in ... - Wiley Online Library

ª Federation of European Neuroscience Societies

European Journal of Neuroscience, Vol. 21, pp. 1733–1740, 2005

Autonomic arousal in an appetitive context in primates: a behavioural and neural analysis Katrin Braesicke,1, John A. Parkinson,1, ,* Yvonna Reekie,1 Mei-See Man,1 Lucy Hopewell,1 Andrew Pears,1 Harriet Crofts,2 Christian R. Schnell3 and Angela C. Roberts1 1

Department of Anatomy, University of Cambridge, Downing Street, Cambridge, CB2 3DY, UK Department of Experimental Psychology, University of Cambridge, Downing Street, Cambridge, UK 3 Angiogenesis Unit ⁄ Oncology, Novartis Institutes for Biomedical Research, Basle, Switzerland 2

Keywords: amygdala, blood pressure, emotion, marmoset, positive affect

Abstract Central to many emotional responses is the accompanying peripheral somatic and autonomic arousal, feedback from which has been hypothesized to enhance emotional memory and to contribute to appraisal processes and decision making, and dysfunction of which may contribute to antisocial behaviour. Whilst peripheral arousal may accompany both positive and negative emotional contexts, its relationship with the former is poorly understood, as are the neural mechanisms underlying such a relationship. The purpose of the present study was to determine the autonomic correlates of anticipation, as well as consumption, of high incentive food, in the freely moving common marmoset and to investigate the contribution of the amygdala to such effects. Blood pressure (BP) and heart rate (HR) were measured remotely by a telemetric device implanted into the descending aorta and behavioural responses were monitored whilst marmosets viewed preferred or non-preferred foods and were then allowed access to eat those foods. A marked rise in blood pressure in unrestrained marmosets was observed in response both to the sight of highly preferred foods (anticipatory period) as well as during the actual consumption of those foods (consummatory period). Excitotoxic lesions of the amygdala abolished the autonomic arousal in the anticipatory period, but spared both the behavioural arousal in the anticipatory period and the autonomic arousal in the consummatory period. Together these data serve as an important step towards understanding the role of autonomic arousal in emotion and its neural underpinnings.

Introduction Emotion is a highly adaptive and complex state incorporating psychological, physiological and behavioural components. The continuum of emotional responses stretches from unlearnt reflexes and fixed action patterns, through Pavlovian learning in which novel stimuli come to elicit conditioned responses, to instrumental behaviour whereby the organism takes active control of the environment to satisfy motivational and emotional requirements. One important function of emotion is to produce appropriate anticipatory responses to improve chances of survival. Central to any emotional response are the accompanying peripheral somatic and autonomic responses, feed back from which has been hypothesized not only to contribute to the perception of arousal (Schachter & Singer, 1962) but also to engage appraisal processes including the self-perception of emotional state (Ekman et al., 1983), to enhance emotional memories (McGaugh, 2004) and to influence various aspects of decision making (Nauta, 1971; Damasio, 1998; for a review of the role of afferent feedback, see Berntson et al., 2003). Indeed, a failure in such mechanisms may contribute to the affective and decision making deficits and social dysfunction in patients with a variety of neuropsychiatric disorders Correspondence: Dr Angela C. Roberts, as above. E-mail: [email protected] *Present address: School of Psychology, University of Wales, Bangor, Adeilad Brigantia, Penrallt Road, Bangor, Gwynedd, LL57 2AS, UK. K.B. and J.A.P. contributed equally to this work. Received 29 November 2004, revised 11 January 2005, accepted 12 January 2005

doi:10.1111/j.1460-9568.2005.03987.x

including dementia of the frontal lobe type (Rahman et al., 2001), manic depression (Hutton et al., 2002) and autism (Hirstein et al., 2001). Considerable progress has been made in our understanding of the behavioural mechanisms underlying emotion and their neural substrates, e.g. (Rolls, 1998; Davis, 2000; Dolan, 2000; Everitt et al., 2000; Gallagher, 2000; LeDoux, 2000), but far less is known of the relationship between somatic and autonomic changes and emotion and their neural underpinnings. However, without an understanding of the interrelationship between brain, behaviour and the periphery it will not be possible to understand fully the neural basis of emotion. A number of brain structures, including the amygdala, orbitofrontal cortex and anterior cingulate cortex have been implicated in autonomic control (Iwata et al., 1986; Neafsey, 1990; Powell et al., 1997; Bechara et al., 1999; Davis, 2000; Mayer et al., 2000; Barbas et al., 2003), but the precise behavioural contexts in which they act are poorly understood, especially with respect to positive affective states. Indeed, on the few occasions that autonomic control in positive affect has been studied, contrasting effects have been observed, e.g. increased blood pressure (BP) in restrained monkeys (Randall et al., 1975) but heart rate (HR) deceleration in unrestrained rats (Hunt & Campbell, 1997), with the effects of neural manipulations remaining unknown. To address these issues, the present study investigated the autonomic correlates of anticipation, as well as consumption, of high incentive food, in the freely moving common marmoset. BP and HR were measured remotely by a telemetric device implanted into the descending aorta (Schnell & Wood, 1993) and behavioural responses

1734 K. Braesicke et al. were monitored whilst marmosets viewed preferred or non-preferred foods (anticipatory period) and were then allowed access to eat those foods (consummatory period). Excitotoxic lesions of the amygdala were performed to determine whether a similar disruption in autonomic arousal would occur in an appetitive context as has been shown previously in negative, fear-inducing contexts.

Materials and methods Subjects Ten common marmosets (Callithrix jacchus; six females and four males) of mean age 23 months at the outset of testing, participated in experiments 1 and 2. All procedures were conducted in accordance with the United Kingdom 1986 Animals (Scientific Procedures) Act under project license PPL 80 ⁄ 1344. Marmosets were not food- or water-deprived prior to, or during, testing as such treatment can change baseline autonomic responses (Young & Landsberg, 1977b).

Apparatus The subjects were tested in an automated apparatus specifically designed for the marmoset (Roberts et al., 1988). A modification to this apparatus incorporated an opening at the rear of the test box providing access to a recessed cylindrical food box (with an internal diameter of 52 mm and length of 51 mm) that was illuminated inside by a 28 V, 0.04 watt encased ceiling light. Two Perspex doors (one opaque and one transparent) formed a barrier between the marmoset and the food box. Opening of the opaque door alone gave animals a view into the food box, whilst access to the food was dependent upon the opening of both doors. Cameras mounted on the inside walls of the chamber recorded the behaviour of the subject and a computer in the adjacent room controlled the doors and other aspects of the behavioural procedure.

Telemetry The telemetry system (Data Sciences, Inc., St Paul, Minnesota) consisted of five basic components: an implantable transmitter (TA11PA-C40) which continuously transmitted the blood pressure of subjects; a receiver (RPC-1) located underneath the behavioural testing box; a calibrated pressure output adapter (R11CPA) with an ambient pressure reference monitor (APR-1) to convert the absolute pressure measured by telemetry to a gauge pressure in millimetres of mercury; an analogue-digital converter [micro 1401, Cambridge Electronic Design (CED)] and a computer based data acquisition system for collection, analysis and storage of the accumulated data (PC with the software package Spike2 Version 2.28, CED).

Surgical procedures For implantation of telemetry probes, marmosets were anaesthetized with a combination of an injection of ketamine sulphate (0.05 mL, i.m.; Pharmacia and Upjohn, Crawley, UK) followed by an injection of Saffan (0.4 mL of a 12 mg ⁄ mL solution, i.m.; Schering-Plough, Welwyn Garden City, UK) and then received an injection of atropine sulphate (0.1 mL of a 600 lg ⁄ mL solution, s.c.; Animalcare Ltd, York, UK). Blood pressure transmitters were placed within the abdomen and the catheter implanted into the descending aorta following procedures described previously (Schnell & Wood, 1993). Non-steroidal analgesics were given on the day of surgery and the day after and antibiotics were given for two days prior to, and up to ten

days following surgery. Marmosets had a two-week recovery period before testing began. For excitotoxic lesions of the amygdala, all marmosets were anaesthetized with a combination of an injection of 0.05 mL ketamine sulphate (Pharmacia and Upjohn, Crawley, UK) i.m. followed by 0.4 mL Saffan (Schering-Plough, Welwyn Garden City, UK) i.m. and maintained with supplementary doses of 0.3 mL Saffan for the duration of surgery. Injections of excitotoxin (0.09 m quinolinic acid, Sigma-Aldrich, UK) were made bilaterally into the amygdala at the following co-ordinates from the interaural line: AP +9.3, L ± 5.6, DV +4.0 and +5.0 (0.2 lL per injection site). Sham-operated controls underwent the same surgical procedure as lesioned animals with the exception that they received infusions of sterile phosphate buffer vehicle rather than excitotoxin. For all placements, infusions were made over 100 s through a stainless steel cannula (30 gauge) attached to a 2 lL precision Hamilton sampling syringe (Precision Sampling Co., Baton Rouge, LA). The cannula then remained in place for 4 min before being withdrawn slowly. Following surgery, all animals were administered 5 mL glucose and saline solution (0.9% saline, 1% sucrose) i.p., followed by Valium (Roche, Hertfordshire, UK) in the range of 0.05 mL to 0.25 mL i.m. intermittently over the first 24 h to suppress any epileptic seizure activity. Animals had a 7–10-day postoperative recovery period before being returned to the test apparatus.

Post-mortem assessment of the excitotoxic lesion of the amygdala All marmosets were given an overdose of anaesthetic (Euthatal; 1 mL of a 200 mg/mL solution, i.p., vericore Ltd, Dundee, UK) and were perfused transcardially with 500 mL 0.1 m phosphate buffered saline, pH 7.4, followed by 500 mL 0.4% paraformaldehyde fixative administered over approximately 10 min. The entire brain was removed and placed in fixative solution overnight before being transferred to a 30% sucrose solution for a minimum of 48 h before sectioning. For verification of lesions, coronal sections (60 lm) of the brain were cut using a freezing microtome and cell bodies stained using Cresyl Fast Violet.

Behavioural procedures Experiment 1 Prior to telemetry implantation, subjects (n ¼ 6) were habituated to the apparatus by pre-exposing the sound and sight of the opening of the doors and the lighting change that would accompany the anticipatory and consummatory periods (Fig. 1a). During these sessions, a small amount of food (several pieces of marshmallow) was available in the food box. Animals were then introduced to the behavioural procedure in which 1–3 anticipatory periods, in which subjects could view the food within the food box, were presented each session (Fig. 1b). For four animals telemetry implantation took place prior to the start of the behavioural procedure allowing for a detailed analysis of behaviour and autonomic activity across acquisition. For the remaining two animals, telemetry implantation took place within the first ten sessions of training. At the start of each anticipatory period the opaque door opened, the houselight was turned off and food box light was illuminated. The duration of each anticipatory period was 20 s and the intervals between periods were pseudo-randomly varied between 40 and 110 s with an average interval of 90 s. The final anticipatory period of a session ended with the transparent door opening allowing 5 min access to the food (consummatory period). In approximately two-thirds of sessions an individual marmoset’s

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 21, 1733–1740

Autonomic arousal in the primate 1735 received training on the task until they too had reached stable autonomic and behavioural performance (ranging from between 16 and 23 sessions), as described in experiment 1 above. Then, after receiving a month break and a further ten control sessions, these marmosets, along with four of the originals received either bilateral excitotoxic lesions of the amygdala (two of the original marmosets and two of the additional marmosets) or a sham control procedure and after a two-week recovery period, animals were re-tested on the same ten sessions that they had received prior to surgery.

Analysis of autonomic activity

Fig. 1. (A) A schematic diagram of the apparatus illustrating the separation of the food box from the rest of the chamber by the transparent door (grey line) and opaque door (black line). The task procedure for each session is summarized in (B). (C) Mean changes in systolic BP (± SEM) during the anticipatory period (compared to the immediately preceding baseline) when viewing preferred (Pref) and non-preferred (Non-pref) foods during the final ten sessions of stable performance (n ¼ 6). (D) Mean duration of food box directed behaviour (± SEM) in the anticipatory period when viewing preferred compared to non-preferred foods and (E) mean difference in systolic BP (± SEM) between the consummatory period and the preceding anticipatory period for sessions with preferred and non-preferred foods. *P ¼ 0.007 and *P ¼ 0.01, comparing preferred and non-preferred foods during the anticipatory period and subsequent consummatory period respectively. vISI, variable interstimulus interval.

preferred foods were presented (foods that were nearly all familiar, e.g. malt loaf, rusks, grapes, marshmallow, raisins, but which differed in their overall level of familiarity) while for the remaining one-third, standard laboratory pellets (non-preferred food) were presented. Whilst animals nearly always ate the preferred foods for the entire 5 min duration of the consummatory period, animals very rarely ate at all during the consummatory period for low incentive pellets. Sessions lasted up to 10 min. Individual subjects were tested until they had reached stable performance on the preferred food sessions, i.e. a change in blood pressure during the anticipatory period that did not differ across five consecutive sessions by more than 20%. Data from these preferred sessions (10 anticipatory periods) along with the adjacent non-preferred sessions (4 sessions and 8 anticipatory periods) were then analysed. In the case of four out of the six marmosets in experiment 1, their stable performance across preferred food sessions was compared with the data from the first five preferred food acquisition sessions (12 anticipatory periods). The number of preferred and non-preferred food sessions that each animal was exposed to before reaching stable performance ranged from between 13 (25 anticipatory periods) to 19 (37 anticipatory periods). Experiment 2 Following a month break, four of the original six marmosets were given a further ten control sessions of training involving 24 presentations of the sight of the food, half of which included preferred and half non-preferred foods. (The signal from the telemetry probes waned in the other two marmosets, such that they could not be continued in the study). Four additional marmosets were implanted with probes and

For the telemetric data, recording failures and outliers were removed (typically blood pressure values above 400 mmHg or under 0 mmHg or other abnormal spikes) and then the systolic and diastolic blood pressure events were extracted as local minima or maxima during one heart beat cycle. The heart rate was calculated using the time interval between the systolic blood pressure events. Outliers or missing values were filled with cubic spline interpolation, though any disruptions in the trace longer than 0.4 s were treated as missing values in the resulting dataset. A mean value was calculated over the 20-s anticipatory period (or from the point when the subject first looked into the food box, for the first anticipatory period of a session) for each autonomic measure (systolic BP, diastolic BP and HR). The immediate 20 s preceding each anticipatory period served as its baseline for comparison purposes. The value for the consummatory period was calculated over the 60 s after the animal had reached into the food box in the case that this happened in the first 20 s after access, or alternatively, the 60 s directly after access when no approach was made within the first 20 s.

Analysis of behaviour For the behavioural data a person blind to experimental conditions scored the videotapes in slow motion with the aid of a clock-timer synchronized to the conditioning programme. The duration of all behaviour directed towards the food box was scored during the 20 s anticipatory periods producing a quantitative measure of ‘food box directed’ behaviour. Major components of the behaviour included orienting towards, ‘looking’ at, and approaching the food box and attempts to reach into the food box, including scrabbling and gnawing at the window. For the consummatory periods the latency for the animal to reach into the food box and the duration of eating over the first subsequent minute and over the entire 5 min were recorded. A general movement measure (duration of locomotion and ballistic limb movement) was also determined for the 20 s baseline period and the subsequent anticipatory period.

Statistical analysis All data were analysed using SPSS version 9. Systolic and diastolic BP, heart rate and food box directed behaviour during the anticipatory period were analysed in study 1 using a one-way analysis of variance (anova) with preference (preferred, non-preferred) as the repeated, within-subjects factor for the performance data. A two-way anova with preference and presence ⁄ absence of reward as the repeated, within subject factors were used when analysing the data from the consummatory period. A paired samples t-test was used to study the acquisition of conditioning. In study 2, all data was analysed using a three-way anova with preference and surgery (pre, post) as repeated withinsubjects factors and lesion (control, amygdala) as the between-group


1736 K. Braesicke et al. factor. Significant interactions were further analysed using simple interaction and simple main effects using the Mean Square error term from the original interaction where appropriate (Howell, 1999).

Results Experiment 1: Characterization of the autonomic correlates of anticipation and consumption of high incentive food All six marmosets showed a heightened systolic BP response to the sight of their preferred foods (mean systolic level: baseline, 119 ± 5.9 mmHg; anticipatory period, 124.7 ± 6.6 mmHg). More importantly, they showed a differential BP response to the sight of their preferred, compared to non-preferred food (main effect of Preference: F1,5 ¼ 19.06, P ¼ 0.007; Fig. 1C). This pattern was mirrored in both diastolic BP (Mean level: baseline, 88.9 ± 5.5 mmHg; anticipatory period: 94.2 ± 5.3 mmHg; main effect of Preference: F1,5 ¼ 14.84, P ¼ 0.012) and HR (Mean level: baseline, 324.9 ± 28.6 beats per minute, bpm; anticipatory period: 340.5 ± 15.1 bpm; main effect of Preference; F1,5 ¼ 6.96, P ¼ 0.046) though the systolic BP response showed the greatest consistency and strength of effects. Marmosets also showed a greater amount of food box directed behaviour to the sight of their preferred compared to non-preferred food (Fig. 1D), consistent with the greater arousing and attentional properties of the preferred food. Paired t-tests of each individual animal’s data demonstrated that five out of the six marmosets showed significantly enhanced behaviour to the sight of their preferred foods (Subject (S)1: t ¼ 4.2, P ¼ 0.001; S2: t ¼ 5.7, P ¼ 0.02; S3: t ¼ 2.73, P ¼ 0.018; S4: P ¼ 4.2, P ¼ 0.001; S5: t ¼ 3.6, P ¼ 0.003) Moreover, there was a positive correlation between the increase in general activity (including food box directed activity) and the rise in BP induced by the sight of the preferred food (Pearson’s correlation ¼ 0.383, P ¼ 0.01). This coupling of both the autonomic response and behavioural activity to anticipation of the preferred food raises the question of a causal relationship between the two; the autonomic activity conceivably being a direct result of the behavioural activity. However, such a relationship was not always seen. One marmoset displayed the second largest anticipatory rise in BP but showed no corresponding increase in activity. In fact this one marmoset displayed the complete opposite pattern of behaviour and displayed far greater activity upon presentation of the non-preferred food even though there was no accompanying rise in BP (S6 t ¼ )3.49, P ¼ 0.003). As a consequence of this animal’s contrasting behavioural pattern of activity a significant main effect of Preference in the group analysis of behaviour was precluded (F1,5 ¼ 5.23, P ¼ 0.071). A rise in systolic BP, over and above that produced when viewing the preferred foods, was also observed when animals were eating the preferred foods (Fig. 1E). No such rise was seen in the equivalent period after the anticipatory period when no reward was presented. anova of changes in BP revealed a Reward Presence–Preference interaction (F1,5 ¼ 16.07, P ¼ 0.01). Simple main effects revealed that there was a rise in BP during the consummatory period for preferred, but not non-preferred, food but only when food was presented (F1,5 ¼ 16.59, P ¼ 0.01). No such rise was seen in the equivalent period after the anticipatory period when no reward was presented and indeed as expected BP decreased, returning to baseline (F1,5 ¼ 10.63, P ¼ 0.022). A similar pattern was observed with diastolic BP (Reward Presence–Preference interaction F1,5 ¼ 6.96, P ¼ 0.046) whilst, in contrast, heart rate fell at the end of the anticipatory period for preferred foods even when the foods were being consumed (main effect of Preference – F1,5 ¼ 11.9, P ¼ 0.018). This differential

Table 1. Changes in systolic BP, diastolic BP, heart rate and behaviour

n¼4

Acquisition Stable performance (first five sessions) (final five sessions)

(i) Anticipatory period Systolic BP (mmHg) 1.073 Diastolic BP (mmHg) 0.540 Heart rate (bpm) )2.640 Behaviour (s) 8.384 (ii) Consummatory period Systolic BP (mmHg) 5.200 Diastolic BP (mmHg) 6.143 Heart rate (bpm) 6.198

± ± ± ±

0.419 0.299 3.970 1.424

± 1.286 ± 1.732 ± 8.770

5.041 5.225 22.729 15.681

± ± ± ±

1.366 (P ¼ 0.04)* 1.711 (P ¼ 0.066) 10.994 (P ¼ 0.079) 1.997 (P ¼ 0.061)

3.837 ± 2.62 3.102 ± 2.848 )10.084 ± 14.303

(i) Mean changes in systolic BP, diastolic BP and heart rate (± SEM) and mean duration of food box-directed behaviour (± SEM) during the 20 s anticipatory period (compared to the immediately preceding 20 s baseline) and (ii) mean changes in systolic BP, diastolic BP and heart rate (± SEM) during the first minute of the consummatory period (compared to the immediately preceding anticipatory period), early in acquisition (first five sessions) and at stable performance (final five sessions) for high incentive sessions only. *P < 0.05, comparing stable performance measures with acquisition measures.

response of BP and HR during the consummatory period may reflect the differential contributions of underlying sympathetic and parasympathetic mechanisms to ingestive processes (Young & Landsberg, 1977a; Steffens et al., 1986). It is unlikely though that ingestive processes contributed to the rise in BP associated with eating the preferred foods in the present study as on the few occasions that an animal failed to show a rise in BP during this period (nine sessions in total taken from different animals) eating still occurred (systolic BP and time spent eating in first 60 s for all animals across both acquisition and performance sessions (mean ± SEM): ingestion period with BP rise, BP rise ¼ 7.45 ± 0.63 mmHg, eating time ¼ 50.14 ± 1.8 s; ingestion period without BP rise, BP rise ¼ )2.17 ± 0.87 mmHg, eating time ¼ 46.72 ± 2.8 s). A comparison of the changes in autonomic and behavioural activity in the final five sessions of stable performance (as described in the Materials and methods) with that seen during the first five sessions of testing, for four out of six of the animals highlighted the acquired nature of the autonomic and behavioural responses in the anticipatory period (Table 1). Using paired sample t-tests it was evident that anticipatory autonomic and behavioural activity were substantially greater in the late stage of acquisition (stable performance) than in the early stage (systolic BP: t ¼ 3.471, P ¼ 0.04, diastolic BP: t ¼ 2.832, P ¼ 0.066, heart rate: t ¼ 2.613, P ¼ 0.079, behavioural activity: t ¼ 2.922, P ¼ 0.061). In contrast, the rise in systolic and diastolic BP during the consummatory period was present from the beginning and remained relatively stable throughout (highest t-value of 1.79 for diastolic BP). The anticipatory autonomic responses to high incentive foods, once acquired, remained relatively stable, being present across two subsequent, two week testing periods, i.e. ten days of pre-lesion performance and ten days of post-lesion performance, in animals that went on to be sham-operated controls in experiment 2 (see Fig. 3). Finally, the number of sessions received before reaching stable performance and the mean level of anticipatory behavioural and autonomic activity occurring at stable performance did not differ between those marmosets implanted prior to the start of behavioural testing and those implanted at a later stage.

Experiment 2: Effects of amygdala lesions on autonomic arousal The schematic diagram of the lesions in Fig. 2 illustrate their extent, which incorporated the lateral, basal, accessory basal and central


Autonomic arousal in the primate 1737

Fig. 2. Excitotoxic lesion of the amygdala. Schematic diagram of a series of coronal sections through the marmoset anterior temporal lobe depicting the size and extent of damage following a quinolinic-acid induced lesion of the amygdala in four marmosets [shadings indicate area of cell loss in all (darkest shading), 3, 2 and only 1 (palest shading) marmoset]. Various nuclei within the amygdala including the lateral (L), basal (B), central (C), accessory basal nucleus (AB), cortical (Co) and medial (M) are depicted on the adjacent half brain sections. Scale bar, 1.5 mm.

nuclei of the amygdala. These nuclei were extensively lesioned in all marmosets with sparing of the most anterior sector of the basal and lateral nuclei (see AP 10.4 in Fig. 2) and the more posterior sectors of the lateral nuclei (see AP 7.7 in Fig. 2). The accessory basal nucleus (AB) was more variably damaged with its medial portion spared in the majority of cases. There was no damage to the cortical nucleus but the medial nucleus (M) was damaged in two out of the four marmosets more posterior (see AP 8.6 in Fig. 2). Apart from bilateral damage to the extreme anterior tip of the hippocampus in just one marmoset there was no extra-amygdala damage. Lesioning the amygdala did not affect baseline BP in any of the animals (F < 1). However, whilst pre-surgery, BP showed the expected increase to the sight of an animals preferred foods (relative

to baseline) this rise was significantly blunted in all animals following the amygdala lesion (Fig. 3A). anova comparing the difference in systolic BP between the anticipatory and baseline periods pre- and post-surgery for preferred and non-preferred foods revealed a three– way interaction (F1,6 ¼ 9.8, P ¼ 0.020). Posthoc analyses revealed a simple interaction between Lesion and Preference post-surgery (F1,6 ¼ 25.196, P < 0.005) but not pre-surgery (F < 1) which was shown subsequently to be due to a selective reduction in the BP rise in the lesioned group (F1,6 ¼ 25.064, P < 0.005) compared to controls. This pattern was mirrored in diastolic BP (3–way interaction F1,6 ¼ 15.421, P ¼ 0.008), but not in HR (3–way interaction F < 1). In contrast, the amygdala lesion had no effect on the duration of food box directed behaviour in the presence of the preferred foods


1738 K. Braesicke et al.

Discussion The present results demonstrate that autonomic arousal accompanies the anticipation as well as the consumption of highly preferred foods in freely moving marmosets. The rise in the anticipatory period developed gradually across testing, presumably as a result of the marmosets learning that the sight of food led to eventual access to that food. In contrast the rise in the consummatory period, when marmosets were consuming their preferred foods, was present from the beginning of testing. Once developed, these arousal responses remained stable over a six-week testing period. The motivational nature of these observed changes was highlighted by the lack of arousal when marmosets viewed their staple diet of dried pellets (nonpreferred food). Excitotoxic lesions of the amygdala abolished autonomic arousal in the anticipatory period but left intact both autonomic arousal in the consummatory period and behavioural arousal in the anticipatory period. Thus together, these findings highlight the selective importance of the amygdala in anticipatory autonomic arousal in an appetitive context.

Autonomic arousal in anticipation of high incentive foods

Fig. 3. Autonomic and behavioural arousal for preferred and non-preferred foods during the anticipatory and consummatory periods before and after excitotoxic amygdala lesions (n ¼ 4 marmosets) or a sham control procedure (n ¼ 4 marmosets). (A) Mean changes in systolic BP (± SEM) during the anticipatory period, compared to the preceding baseline period. (B) Mean duration of food box directed behaviour (± SEM) during the anticipatory period. (C) Mean changes in systolic BP (± SEM) during the consummatory period, compared to the immediately preceding anticipatory period. *P ¼ 0.005, group difference.

(Fig. 3B), indicating the relative independence of behavioural and autonomic arousal. In addition, there was no effect of the amygdala lesion on the consummatory-induced rise in BP associated with the ingestion of highly preferred foods (Fig. 3C) and no effects on the amount of time spent eating. anova of food box behaviour pre- and post-surgery (3–way interaction F < 1) revealed no interaction although, as expected, there was a main effect of Preference (F1,6 ¼ 309.11, P < 0.001). Similarly, anova comparing the rise in systolic BP during the consumption period pre- and post-surgery revealed no effect of Lesion (F < 1) and no interactions involving Lesion (all Fs < 1). The significant effect of Preference (F1,6 ¼ 44.373, P ¼ 0.001) confirms that there was a significant rise in BP during consumption of the preferred food, over and above the preceding anticipatory period. As the pattern of autonomic and behavioural responses of all four lesioned marmosets, compared to that of sham controls, was relatively consistent, any differences between the lesioned marmosets in their pattern of amygdala damage, e.g. only two out of four marmosets exhibited damage to the medial nucleus, did not correlate with performance.

Systolic BP showed the most marked and consistent rise during the anticipatory period for high incentive food whilst HR, although showing significant increases to the sight of preferred foods, was not as tightly linked to motivational manipulations and may reflect the influence of other factors including gross movement (e.g. Roberts & Young, 1971; Randall & Smith, 1974). Moreover, HR, unlike systolic and diastolic BP, did not increase during the consumption of the highly preferred foods. The independence of BP changes and behavioural activity was seen most clearly in experiment 2 when amygdala lesions left anticipatory behavioural activity intact whilst abolishing anticipatory autonomic activity. Although this study does not address the underlying peripheral control of the autonomic responses observed, it is not unreasonable to assume that sympathetic activation played a pivotal role in increases in the BP response during the anticipatory and consummatory periods, as has been suggested previously (Randall & Smith, 1974). Such positive anticipatory autonomic arousal not only acts to prepare the body for engaging in activities that will eventually result in the retrieval and consumption of the food itself, but has been proposed to, in addition, feed back on the brain to enhance the overall memory of the event (McGaugh, 2004), potentiate the self perception of emotion (Ekman et al., 1983) and to engage decision making processes (Damasio, 1998). An increase in BP during presentation of a stimulus associated with the subsequent delivery of reward (conditioned stimulus, CS) has been reported previously in restrained monkeys (Randall et al., 1975) but such conditioned arousal has not always been seen (Harris & Brady, 1974; Hunt & Campbell, 1997). However, when conditioned HR deceleration has been reported (Hunt & Campbell, 1997) it has been suggested that this may reflect an attentional (Kapp et al., 1992) rather than a motivational mechanism and consistent with this proposal the HR deceleration has been abolished by parasympathetic but not sympathetic blockade. Indeed a similar bradycardia has been reported to accompany attentional processing in humans (Middleton et al., 1999). Whether or not a sympathetically aroused state is induced in an appetitive context may well depend upon the overall incentive value of the reward. Autonomic arousal occurs when reward value is high, as in the present study in which reward was a large box of a marmoset’s highly preferred food retrieved at the end of some sessions, but not others, and, as in the study of Randall et al., (1975), in which the food reward was the monkey’s daily food ration. In contrast, no such


Autonomic arousal in the primate 1739 arousal was seen when the repeated presentations of two small sucrose pellets acted as reward (Hunt & Campbell, 1997). The sight of the food in the present study may also have been a particularly good inducer of positive arousal because the physical characteristics of the food itself, e.g. its shape and colour, through their close temporal and spatial proximity with the primary rewarding properties of the food, i.e. taste, may form strong associations with these rewarding properties and thus become powerful conditioned reinforcers in their own right (LoLordo, 1979). Amygdala lesions dissociate autonomic from behavioural arousal in anticipation of high incentive food The loss of autonomic arousal produced in anticipation of food reward following excitotoxic lesions of the amygdala in the present study would appear comparable to the attenuation of autonomic arousal produced to fear-inducing conditioned stimuli (CS) in amygdala lesioned rats (Iwata et al., 1986). As the lesion in the present study included both the central nucleus (CN) and the basolateral complex (BLN) it is not possible to specify whether damage to the latter or former was responsible for this effect although damage to the CN would seem the most likely (Iwata et al., 1986). Regardless of which amygdaloid nucleus contributes to the observed autonomic activity these findings highlight the importance of the amygdala in the expression of autonomic arousal in an appetitive context, which complements its proposed role in the expression of autonomic arousal in aversive contexts. Thus it is consistent with the hypothesis that the amygdala can provide information about the intensity of an emotional stimulus irrespective of valence (Anderson et al., 2003; McGaugh, 2004). That this same amygdala lesion did not affect behavioural arousal directed towards the food box highlights the neural fractionation of the different elements of an emotional response. Previously it has been shown that the motivated approach response towards the food hopper during presentation of an appetitive CS is unaffected by excitotoxic lesions of either the basolateral or central nucleus of the amygdala (Holland & Gallagher, 1999), whilst other CS-elicited behaviours are disrupted by these lesions, including autoshaping and CS orienting responses by central nucleus lesions (Cardinal et al., 2002) and second-order conditioning by basolateral nucleus lesions (Hatfield et al., 1996). The current study demonstrates that even after combined lesions of both these nuclei of the amygdala, the behavioural arousal elicited by the sight of high incentive food remains present, despite the loss of autonomic arousal. Whether these amygdala independent, behavioural approach responses to high incentive food are mediated instead by a circuit involving the temporal lobe recognition system and the ventral striatum remains to be determined. What is clear though is that these autonomic and behavioural arousal responses are dissociable implying that autonomic arousal does not contribute directly to this particular behavioural response. However, it does not rule out a role for autonomic arousal in enhancing the overall vigour of such a response, as this was not easily assessed in the present study. Moreover, it does not rule out a role for autonomic arousal in the acquisition, rather than the expression, of such behaviour or, perhaps more likely, in the acquisition and ⁄ or expression of other more complex behaviours involving instrumental goal directed actions in the contexts of conflicting negative and positive outcomes. Another important finding in the present study was the preservation of the autonomic arousal in the consummatory period in comparison to the loss of autonomic arousal in the preceding, anticipatory period of amygdala lesioned animals. This consummatory arousal is unlikely to be due to the sensory processes of ingestion per se as, on occasion,

ingestion of the food was observed to take place in the absence of a BP rise. More likely, the hedonic aspects of ingesting highly preferred foods acted to trigger a BP rise; an effect independent of the amygdala. In previous studies the effects of amygdala lesions on consummatory responses has been examined with respect to fear with results varying from total abolition (Blanchard & Blanchard, 1972; Goldstein et al., 1996) to partial reduction (Bechara et al., 1999) or no reduction at all (Antoniadis & McDonald, 2001). However, in some of these studies, results may have been confounded by the presence of extra-amygdala damage (Blanchard & Blanchard, 1972; Bechara et al., 1999) or by conditioning (Goldstein et al., 1996). The results here are consistent with those of Antoniadis & McDonald (2001) examining unconditioned freezing responses to shock and suggest that in an appetitive context, autonomic arousal accompanying the consummatory responses to high incentive food is amygdala independent. The identity of the neural structures upon which consummatory autonomic arousal is dependent remain to be determined but likely candidates include the shell region of the nucleus accumbens and circuitry within the brainstem, structures already implicated in the control of positive affective behavioural reactions to pleasant sensations such as sweet tastes (Berridge, 2003). In conclusion, this study provides a convergence of behavioural, autonomic and neural analyses of positive emotion in the primate. The sight of highly motivating stimuli increase behavioural and autonomic arousal, the latter being dependent upon the integrity of the amygdala. Future studies will determine the extent to which, and the context in which, other neural structures, including the orbitofrontal and anterior cingulate cortex, control the expression of such positive autonomic arousal. In addition, the way in which autonomic arousal may feed back to influence the vigour or direction of subsequent goal-directed behaviour and the neural basis of any such effect may also be addressed using this approach and will have important implications for our understanding of the interface between emotion and cognition in healthy and psychiatric populations.

Acknowledgements We thank Hugh Middleton for invaluable technical advice, Caroline Parkinson for preparation of histological material and Ros Ridley for supplying marmosets. Y.R and A.P. were supported by BBSRC research studentships. We thank Professors T. W Robbins and B. J Everitt for their helpful comments on the manuscript. The work was supported by a Medical Research Council Career Establishment Grant to ACR and is a publication within the Medical Research Council Centre on Behavioural and Clinical Neuroscience.

Abbreviations BP, blood pressure; CS, conditioned stimulus; HR, heart rate.

References Anderson, A.K., Christoff, K., Stappen, I., Panitz, D., Ghahremani, D.G., Glover, G., Gabrieli, J.D. & Sobel, N. (2003) Dissociated neural representations of intensity and valence in human olfaction. Nature Neurosci., 6, 196–202. Antoniadis, E.A. & McDonald, R.J. (2001) Amygdala, hippocampus, and unconditioned fear. Exp. Brain Res., 138, 200–209. Barbas, H., Saha, S., Rempel-Clower, N. & Ghashghaei, T. (2003) Serial pathways from primate prefrontal cortex to autonomic areas may influence emotional expression. BMC Neurosci., 4, 25. Bechara, A., Damasio, H., Damasio, A.R. & Lee, G.P. (1999) Different contributions of the human amygdala and ventromedial prefrontal cortex to decision-making. J. Neurosci., 19, 5473–5481.


1740 K. Braesicke et al. Berntson, G.G., Sarter, M. & Cacioppo, J.T. (2003) Ascending visceral regulation of cortical affective information processing. Eur. J. Neurosci., 18, 2103–2109. Berridge, K.C. (2003) Pleasures of the brain. Brain Cogn., 52, 106–128. Blanchard, D.C. & Blanchard, R.J. (1972) Innate and conditioned reactions to threat in rats with amygdaloid lesions. J. Comp. Physiol. Psychol., 81, 281– 290. Cardinal, R.N., Parkinson, J.A., Lachenal, G., Halkerston, K.M., Rudarakanchana, N., Hall, J., Morrison, C.H., Howes, S.R., Robbins, T.W. & Everitt, B.J. (2002) Effects of selective excitotoxic lesions of the nucleus accumbens core, anterior cingulate cortex, and central nucleus of the amygdala on autoshaping performance in rats. Behav. Neurosci., 116, 553–567. Damasio, A.R. (1998) The somatic marker hypothesis and the possible functions of the prefrontal cortex. In Roberts, A.C., Robbins, T.W. & Weiskrantz, L., (Eds), Prefrontal Cortex: Cognitive and Executive Functions, Oxford University Press, Oxford, pp. 36–50. Davis, M. (2000) The role of the amygdala in conditioned and unconditioned fear and anxiety. In Aggleton, J.P., (Ed), The Amygdala, Oxford University Press, New York, pp. 213–288. Dolan, R.J. (2000) Functional neuroimaging of the amygdala during emotional processing and learning. In Aggleton, J.P. (Ed), The Amygdala, Oxford University Press, New York, pp. 631–654. Ekman, P., Levenson, R.W. & Friesen, W.V. (1983) Autonomic nervous system activity distinguishes among emotions. Science, 221, 1208–1210. Everitt, B.J., Cardinal, R.N., Hall, J., Parkinson, J.A. & Robbins, T.W. (2000) In Aggleton, J.P., (Ed), The Amygdala, Oxford University Press, New York, pp. 353–390. Gallagher, M. (2000) The amygdala and associative learning. In Aggleton, J.P. (Ed), The Amygdala, Oxford University Press, New York, pp. 311–330. Goldstein, L.E., Rasmusson, A.M., Bunney, B.S. & Roth, R.H. (1996) Role of the amygdala in the coordination of behavioral, neuroendocrine, and prefrontal cortical monoamine responses to psychological stress in the rat. J. Neurosci., 16, 4787–4798. Harris, A.H. & Brady, J.V. (1974) Animal learning. Visceral and autonomic conditioning. Annu. Rev. Psychol., 25, 107–133. Hatfield, T., Han, J.S., Conley, M., Gallagher, M. & Holland, P. (1996) Neurotoxic lesions of basolateral, but not central, amygdala interfere with Pavlovian second-order conditioning and reinforcer devaluation effects. J. Neurosci., 16, 5256–5265. Hirstein, W., Iversen, P. & Ramachandran, V.S. (2001) Autonomic responses of autistic children to people and objects. Proc. R. Soc. Lond. B Biol. Sci., 268, 1883–1888. Holland, P.C. & Gallagher, M. (1999) Amygdala circuitry in attentional and representational processes. Trends Cogn. Sci., 3, 65–73. Howell, D.C. (1999) Statistical Methods for Psychology Duxberry Press, Pacific Grove, USA. Hunt, P.S. & Campbell, B.A. (1997) Autonomic and behavioral correlates of appetitive conditioning in rats. Behav. Neurosci., 111, 494–502. Hutton, S.B., Murphy, F.C., Joyce, E.M., Rogers, R.D., Cuthbert, I., Barnes, T.R., McKenna, P.J., Sahakian, B.J. & Robbins, T.W. (2002) Decision making deficits in patients with first-episode and chronic schizophrenia. Schizophr. Res., 55, 249–257. Iwata, J., LeDoux, J.E., Meeley, M.P., Arneric, S. & Reis, D.J. (1986) Intrinsic neurons in the amygdaloid field projected to by the medial geniculate body mediate emotional responses conditioned to acoustic stimuli. Brain Res., 383, 195–214.

Kapp, B.S., Whalen, P.J., Supple, W.F. & Pascoe, J.P. (1992) Amygdaloid contributions to conditioned arousal and sensory information processing. In Aggleton, J.P. (Ed), The Amygdala: Neurobiological Aspects of Emotion, Memory and Mental Dysfunction, Wiley-Liss, New York, pp. 229–254. LeDoux, J. (2000) The amygdala and emotion: a view through fear. In Aggleton, J.P. (Ed), The Amygdala. Oxford University Press, New York, pp. 289–310. LoLordo, V.M. (1979) Selective associations. In Boakes, A.D.R.A. (Ed), Mechanisms of learning and motivation. Lawrence Erlbaum Associates, Hillsdale, pp. 367–398. Mayer, E.A., Naliboff, B. & Munakata, J. (2000) The evolving neurobiology of gut feelings. Prog. Brain Res., 122, 195–206. McGaugh, J.L. (2004) The amygdala modulates the consolidation of memories of emotionally arousing experiences. Annu. Rev. Neurosci., 27, 1–28. Middleton, H.C., Sharma, A., Agouzoul, D., Sahakian, B.J. & Robbins, T.W. (1999) Contrasts between the cardiovascular concomitants of tests of planning and attention. Psychophysiology, 36, 610–618. Nauta, W.J. (1971) The problem of the frontal lobe: a reinterpretation. J. Psychiatr. Res., 8, 167–187. Neafsey, E.J. (1990) Prefrontal cortical control of the autonomic nervous system: anatomical and physiological observations. Prog. Brain Res., 85, 147–165; discussion 165–146. Powell, D.A., Chachich, M., Murphy, V., McLaughlin, J., Tebbutt, D. & Buchanan, S.L. (1997) Amygdala–prefrontal interactions and conditioned bradycardia in the rabbit. Behav. Neurosci., 111, 1056–1074. Rahman, S.B.J.S.R.N.C., Rogers, R. & Robbins, T. (2001) Decision making and neuropsychiatry. Trends Cogn. Sci., 5, 271–277. Randall, D.C., Brady, J.V. & Martin, K.H. (1975) Cardiovascular dynamics during classical appetitive and aversive conditioning in laboratory primates. Pavlov. J. Biol. Sci., 10, 66–75. Randall, D.C. & Smith, O.A. (1974) Ventricular contractility during controlled exercise and emotion in the primate. Am. J. Physiol., 226, 1051–1059. Roberts, A.C., Robbins, T.W. & Everitt, B.J. (1988) The effects of intradimensional and extradimensional shifts on visual discrimination learning in humans and non-human primates. Q. J. Exp. Psychol. B, 40, 321–341. Roberts, L.E. & Young, R. (1971) Electrodermal responses are independent of movement during aversive conditioning in rats, but heart rate is not. J. Comp. Physiol. Psychol., 77, 495–512. Rolls, E. (1998) Neurophysiology and functions of the primate amygdala and the neural basis of emotion. In Roberts, A.C., Robbins, T.W. & Weiskrantz, L. (Eds), Prefrontal Cortex: Cognitive and Executive Functions, Oxford University Press, Oxford, pp. 67–86. Schachter, S. & Singer, J.E. (1962) Cognitive, social, and physiological determinants of emotional state. Psychol. Rev., 69, 379–399. Schnell, C.R. & Wood, J.M. (1993) Measurement of blood pressure and heart rate by telemetry in conscious, unrestrained marmosets. Am. J. Physiol., 264, H1509–H1516. Steffens, A.B., Van der Gugten, J., Godeke, J., Luiten, P.G. & Strubbe, J.H. (1986) Meal-induced increases in parasympathetic and sympathetic activity elicit simultaneous rises in plasma insulin and free fatty acids. Physiol. Behav., 37, 119–122. Young, J.B. & Landsberg, L. (1977a) Stimulation of the sympathetic nervous system during sucrose feeding. Nature, 269, 615–617. Young, J.B. & Landsberg, L. (1977b) Suppression of sympathetic nervous system during fasting. Science, 196, 1473–1475.
