Conference 2002

Workshop I
Workshop II
Workshop III
Workshop IV
Final Group Discussion and Summary
Schedule

Photos of the June 2002 Meeting

Workshop III: Chemical Signaling
Chair: P. O'Donnell; Scribe: A. Hariri & M. Marinelli; Spokesperson; M. Marinelli

Most of our discussion was centered on dopamine (DA). No matter how hard (sort of) we tried to depart from DA, the conversation kept on going back to dopamine.
We mainly focused on the first two questions, but we tried to cover all five questions.

1. Do emotionally negative stimuli activate or suppress DA neurons?
Overall, we concluded "activate"!

We realized that it was difficult to answer this question because most studies looking at the role of DA in behavior have focused on its role on positive "rewarding" stimuli; the labs that study DA cell activity are mostly interested in looking at the role of DA in reward, and not in response to emotionally negative stimuli. For example, we realized that although all of us knew that neuroleptics are "reward-blockers", none of us knew their effects in fear-conditioning paradigms. These studies may exist, but we were not aware of them. Since then, we have looked things up and realized that, in fact, DA blockade (esp. of the mesoamygdala pathway) does impair the retrieval of conditioned fear associations (work from several labs like Le Doux, Kokkinidis, Nagase). Therefore DA is important in the processing of fear conditioning and possibly emotionally-negative stimuli.

Initially there seemed to be a few controversies on whether DA transmission is activated or suppressed following emotionally-negative stimuli. Microdialysis data show that Nucleus Accumbens (NAc) DA is increased (preferentially in the shell) following stressful stimuli (e.g. Kalivas & Duffy, 1995; Barrot et al. 1999; Wu et al., 1999). DA is also increased preferentially in the PFC (e.g. Thierry et al., 1976; Abercrombie et al., 1989). However, electrophysiological recordings in monkeys (Schultz, several papers) show that whereas positive unexpected stimuli activate midbrain DA cells, the absence of an expected reward can actually decrease DA cell activity. In addition exposure to an "aversive" stimulus does not modify DA cell activity. It was noted, however, that the intensity of the negative stimuli was very mild in these studies (air puffs) and did not correspond to the intensity of the positive stimuli. It is therefore possible that more emotionally negative stimuli could produce the activation of DA cells. In fact, Moore, Rose and Grace (2001) have shown that stress (cold stress) increases the number of cells exhibiting high levels of bursting activity (>50% bursting), and Marinelli, Cooper & White (unpublished) have also shown that stress (mild food restriction) increases the impulse activity of midbrain DA cells.

We then sought to further the understanding of DA cell "function" in the processing of aversive/positive stimuli by looking at the anatomical outputs/inputs from and to midbrain DA cells. DA cells receive innervations from the superior colliculus. Although this structure processes visual information, the deeper layers are involved with multi-sensory integration. In addition, DA cells receive input from the PFC and amygdala. There were some controversies as to whether the central nucleus of the amygdala (CeA) projects to the VTA or not. It was brought up that work from Zham's group states that the CeA projects to the substantia nigra pars compacta, and the fibers only pass through the VTA, without innervating it. On the other hand, work by Julie Fudge & Suzanne Haber shows that the CeA does indeed innervate the VTA. DA cells also largely project to the NAc, as well as to the PFC, and amygdala (it was mentioned that work from Swanson's group has shown that each neuron projects selectively to one structure). Activation of DA cells could, therefore, be translated into a generalized activation/inhibition of these structures (although preferential activation of certain pathways can be achieved in certain conditions - e.g. stress preferentially increases DA in the PFC than in the striatum; drugs preferentially increase DA in the NAc). Given these inputs/outputs, it was suggested that the information from and to these structures could impart a "value" on the stimulus that is received by the DA cells. In other words, DA could serve as a signal but it would be the inputs/outputs that provide a "label" to the stimulus. DA cells projecting to the different structures are activated, and the receiving structures respond and integrate the received signal.

2. Which aspects of a stimulus drive the DA system (novelty, reward, salience)?
Our initial conclusion was "novelty and salience"; then we realized that reward is often perceived as a novel stimulus, so our final conclusion was that DA is driven by all three: novelty, reward, and salience.

Our discussion started off in a way that was very similar to the first few paragraph of the recent Kelley & Berridge review in JNS (2002). Does DA mediate reward, pleasure, salience, learning or, approach; does it increase signal/noise, does it "prepare" the brain to receive a stimulus?

We found little direct evidence for any of these hypotheses. But we also found little direct evidence against them. A typical example concerned the hypothesis regarding signal/noise ratio; we realized that although there is no strong evidence for DA increasing this ratio, there is also no evidence contradicting it.

We questioned whether a change in DA levels, DA release, or DA cell activity mediates the effects of DA and we questioned the "location" of DA action (PFC, NAc, Striatum).
Concerning the possibility that DA cell activity mediates reward, pleasure, salience etc… we mentioned that DA cells have a pacemaker activity. So it is possible that changes in this activity (caused by stimuli or drugs) may be a signal that "prepares" the brain to receive a stimulus. However it was mentioned that it is a simplistic view, beautiful, but too simple to believe that and increase in DA just prepares the brain for inputs.

This led us to refining the idea that DA "prepares" the brain to one where DA could determine "arousal". We therefore asked whether DA can modify "attention", and maybe even sleep/wake patterns. However, as soon as that was said, it was stated that the term "arousal" has a very lousy history in psychology because it is so vague. And we thought that we should not associate DA with such a vague notion.

We thus departed from the notion that DA could set a level of arousal, and discussed the possibility that DA could have differential effects according different states. For example, a child that is relaxed will turn towards a novel stimulus, whereas the same child will avoid the novel stimulus if it is stressed. It is therefore likely that DA could have different effects on the individual according to the state an individual is in. We could extrapolate this to actual cells. DA could have state-dependent effects of cell activity. Indeed, it has been shown that DA receptor activation can be either excitatory or inhibitory on post-synaptic cells (PFC or NAc) whether the cell is in an excited or depressed state (published work by Patricio O'Donnell, and unpublished studies by Don Cooper & Frank White).

We then mentioned that DA could play a role in approach/avoidance. Or possibly that certain aspects of DA transmission could determine the switching from one behavior to the other. In particular, the passing from tonic to phasic burst firing might serve as a switching signal that facilitates mental act and switch if behaviors. The fact that DA could serve as a switching signal was partly inspired by Redgrave's theory on this subject. Interestingly, patients with Parkinson's disease show deficient sensitivity in task switching.

We talked extensively about Schultz's data again. DA activity increases with expected rewards and a cue will always increase DA cell activity. We wondered how DA cells could be so "intelligent" as to calculate the time when the reward should come. It is possible that feedback to DA neurons through limbic structures might inform DA cells about what is happening; through these inputs the cells could "learn" about events and their value. We then talked about the outputs, and how these could also facilitate learning of events. However we mentioned that "memory" largely resides in the hippocampus, but this structure does not receive strong DA innervations.

We also talked about individual differences in DA cell activity, and whether these might be related to increased risk for emotion-related responding. It was shown (Marinelli & White, 2000) that there are indeed individual differences in DA cell activity, and animals with enhanced impulse activity have increased vulnerability to develop cocaine self-administration. Other situations that increase drug vulnerability (e.g. repeated drug treatment and stress) also increase DA cell activity, further suggesting that the activity of these cells could play an important role in facilitating addictive behaviors, or the learning of reward-related tasks.

If increased activity of DA cells increases reward-related behaviors, we wondered whether DA cells of "vulnerable" individuals receive more excitatory inputs. Excitatory input to DA cells could come from the PFC. However it was recently shown by Bita Moghaddam's group that stimulation of PFC at physiologically-relevant frequencies (rather than non-physiological frequencies), decreases, rather than increases DA concentrations in the NAc. So we wondered where the excitatory input to DA cells comes from. In addition, if it does come from the PFC, it was shown by Susan Sesack's group that the DA cells that receive PFC inputs do not project to the NAc; they only project back to the PFC. It was mentioned that DA cells have been shown to have gap junctions, so it is possible that, although excitatory input does not reach NAc-projecting cells, the presence of gap junctions could propagate the excitatory signal throughout neighboring DA cells, and thus reach those cells that do project to the NAc.

3. Does NE and/or 5HT signaling have an emotional component (intensity, valence, salience)?
We somehow managed to go back on DA… if these other neurotransmitters have such a role, it is because they modulate DA transmission!

We were convinced that other monoamines definitely play a role in signaling emotion, and that an interplay of catecholamines might be important in modulating emotion and motivation. However, as soon as we said that, it was stated that NE, through a1 receptors, modulates DA cell activity. So, we went back to talking about DA…

The anatomists in our group brought a lot of insight as to the possible role of other neurotransmitters in emotion/motivation. Given the anatomical distribution catecholamine fibers and receptors, we gained some insight as to their possible role. For example, 5-HT has a very widespread distribution, and is very diffuse in cortical areas (instead DA is mainly present in limbic and motor areas, and PFC). So it is possible that 5-HT could, this time, really serve to induce a general state of "arousal" or attention. To back this up, it was pointed out that 5-HT cells show a rather tonic activity (with little bursting) and they are mostly inactive during sleep, they show slow firing during wake cycles, and become fast firing during wake cycles when the animal is moving.

Concerning NE, we concluded that NE in might play an important role in the consolidation of memory. In fact, blockade of NE transmission in the PFC impairs the ability to retrieve learned tasks. No effects are seen when manipulating NE transmission in the hippocampus, attributing a major role to PFC NE in memory consolidation.

4. What is the contribution of CRF/HPA axis activation to appetitive versus aversive motivational states?
We concluded that there is an inverted U-shaped curve concerning the effects of HPA-axis activation in the modulation of motivational states. "Pleasure always comes with a little bit of pain…"

It is undoubtedly true that stress and glucocorticoid secretion can both activate and inhibit learning and reward-related tasks. In fact, the influence of these hormones on behavior has an inverted U-shaped curve; low doses facilitate learning, whereas high doses impair learning and also increase hippocampal cell death. The same inverted U-shaped curve is also seen for the effects of corticosterone on relapse of drug-seeking behavior. Similarly, mild stress (which slightly increases glucocorticoid levels) can facilitate attention, learning and drug effects, whereas chronic severe stress (which produces massive elevation of glucocorticoids) has opposite effects.

The inverted U-shaped curve of stress on behavior is strikingly similar to the inverted U-shaped curve of stress on DA levels. In fact, stress increases NAc DA, but chronic or very intense stress actually decreases DA levels.

It was mentioned that glucocorticoids modulate DA levels, and it is interesting to note that corticosterone is self-administered both orally and intravenously in rats (possibly because it increases DA?). In humans, cortisol produces elation, and excitation, and is often associated with antidepressant treatment to improve outcome of antidepressant therapy in depressed patients. In addition, individuals sometimes "self-administer" stress; these same individuals that seek stressful situations are actually those that are more at risk for developing drug self-administration. So there is a tight connection between stress and the activation of reward pathways. As said before, pleasure always comes with a little bit of pain.

We also mentioned that the role of the HPA axis is rather complex; for example CRF is anxiogenic if injected in the brain, even though it is not released in the brain.

5. What is the contribution of other neurotransmitters or neuromodulators? - opioids and other peptides; acetylcholine.
We first said "complicated" and then decided that "complex" was more appropriate. We somehow ended up with the notion that these "other" neurotransmitters can play a role in modulating emotions, but… essentially through an action on DA!

Unfortunately, I have enormous gaps in my notes. Maybe Ahmad Hariri, who was the scriber in our session, might have better notes.

We mentioned that nicotine is neuroprotective on DA cells; so we sort of re-drifted back on DA.

On this line, we talked about data showing that opioid receptor activation increases DA cell activity and DA release. However, it was mentioned that many opioid-dependent behaviors can be maintained in the absence of an intact DA transmission; thus µ receptors can mediate "reward" in a DA-independent manner.

We also mentioned that opioid receptor concentrations in the amygdala are inversely related with amygdala response in PET studies, suggesting that opioids could "dull" the fear response.

Workshop III: Chemical Signaling
Chair: A. Kelley; Scribe: C. Bradberry; Spokesperson; S. Hendricks

We focused our discussion on norepinephrine (NE) and serotonin (5-HT) from the beginning, with the general consensus that these neurotransmitters are not considered in the neurobiology of motivation and emotion.as much as they deserve to be. As regards 5-HT, the strong link of reduced indices of 5-HT function to impulsivity was thought to be very important. This link is seen both clinically and preclinically. The most extensive literature is in the area of observed correlations between impulsivity and CSF levels of the 5-HT metabolite 5-HIAA. It can also be shown in rats that lesion of cortical 5-HT increases impulsivity on a go/no go task. This is not a general deficit in cognitive function, as these animals actually learn visual conditioned discriminations faster than controls. However, the increased impulsivity associated with reduced 5-HT cannot be linked with a particular motivational state.

5-HT/dopamine interactions were mentioned, which brought up a question as to the similarity of drug dependence and impulsivity. In this regard, the confound of potential effects of drug exposure on serotonergic systems was recognized. Also, 5-HT receptor knockout mice have not presented a clear answer here, perhaps in part due to pre- vs. post-syanptic differences in receptor location, and the generally very complex nature of serotonin receptorology.

The influence of NE on 5-HT function was noted - for example, a tonic level of alpha-1 activation must be provided for electrophysiological study of 5-HT neuronal activity in-vitro. There has been less interest in in-vivo studies. The question was raised as to whether evidence of increased 5-HT release may result from recruitment of "silent neurons."

Serotonin selective reuptake inhibitors (SSRIs) were discussed as one way of increasing 5-HT tone, though perhaps not to a large extent due to feedback regulation. Here the point was raised as to a potential role of increased 5-HT in anxiety. The less effective short version of the 5-HT transporter linked polymorphism has been shown to be associated with anxiety, and that may be due to increased synaptic levels of 5-HT.

Here the discussion turned to the use of SSRIs in the treatment of obsessive compulsive disorder (OCD). It was noted that a mixed action on both NE and 5-HT is necessary for effectiveness. When they are effective, there are alterations in cortical blood flow. OCD compulsions have a cognitive aspect, but they are also characterized by a highly motivated state. Is this a cortical disorder? Cortical activation is not seen in symptom provocation studies. What is cognition and what is emotion? The impetus for cognitive thought has a large affective component. Could there be a lack of control from the prefrontal cortex? It was proposed that cognitive therapy might differ in route of effect from SSRIs, perhaps via alterations in the amygdala.

In general, the role of NE in behavior and arousal is more clear than for 5-HT. There is an inverted-U dose effect curve for NE that can result in a narrow optimal dose range. (It was suggested that DA and NE are the yin and yang of the CNS - one makes you feel good and the other makes you feel anxious.) NE depletions impair learning that requires attending to discrete stimuli, but improve learning based on contextual cues.

Anatomically, NE and DA tend to avoid each other, but where they do overlap is often in affect oriented areas such as the amygdala and ventral frontal cortex.

Is there a role for the nucleus accumbens in fear and feeding? Infusion of muscimol in the anterior shell induces feeding and conditioned place preference. In more posterior regions, it causes more anxiety and fear-like behavior. Posterior nucleus accumbens also has more corticotropin releasing factor (CRF) and NE and is also approaching the bed nucleus of the stria terminalis (BNST).

Are there cells hard-wired to approach/avoidance behavior in one integrated structure? This possibility raised a concern about the "extended amygdala" concept because of a long association of the BNST with fear and anxiety. In the BNST, CRF mediates stress-induced relapse to cocaine and heroin self-administration. There was also a discussion about the relative importance of CRF projections to the locus coeruleus from central nucleus of the amygdala vs from the BNST.

There was discussion of DA opioid interactions. Opioid infusion can increase DA release, but some effects are not DA dependent. For example, opioid infusion into the accumbens induces feeding behavior that is not blocked by DA antagonists. Also, DA antagonism in the nucleus accumbens does not block heroin SA either systemically, or directly into the nucleus accumbens. DA denervation sensitizes the locomotor response to NAC infusion of morphine, and DA antagonism can increase the rewarding effects of morphine and nicotine.