Alvaro Pascual-Leone,

March 20, 2009
11th Annual Lecture

Alvaro Pascual-Leone

 


0

Alvaro Pascual-Leone, MD, PhD
Berenson-Allen Center for Noninvasive Brain Stimulation
Beth Israel Deaconess Medical School
Harvard Medical School

 


Modifying Decision Making

Abstract: In recent years, dual-process theories that contrast automated and controlled processes have been put forward to explain different areas of human cognition. In this context, will-power refers to goal-driven cognitive control or regulation of impulses, passions, cravings, and habits. Such regulation may be conceptualized as cognitive control over the balance between a “cool”, reflective mental system that effortfully represents rational and reasoned goals, such as long-term mental and physical health, and a “hot”, reflexic mental system that automatically guides quick, impulsive, and emotional responses to environmental stimuli.

In recent years, lesion and functional neuroimaging studies suggest that the prefrontal cortex is a critical component of the neural circuitry engaged when people voluntarily and consciously regulate their behavior. In addition to neuroimaging studies, lesion studies suggest that particularly the right prefrontal cortex plays a central role in behavioral regulation and the control of impulsive, reflexic tendencies.

Modulation of will-power and dual-process theories offer a valuable framework that can serve to guide translational insights from cognitive neuroscience into the clinic. Proof-of-principle studies reveal that noninvasive brain stimulation of the dorsolateral prefrontal cortex with repetitive transcranial magnetic stimulation or transcranial direct current stimulation can influence decision-making, enhance will-power and promote reflective processes in healthy subjects. The same type of noninvasive brain stimulation can suppress alcohol, cocaine, nicotine and even food craving in patients, who are known to have impaired decision-making behaviors. Modulation of decision making, and enhanced cognitive regulation of emotion, reward, and gratification could have widespread mental and physical health benefits, including mood disorders, anxiety, ADHD, PTSD, substance abuse, smoking, and obesity.

Speaker Introduction

John Trueswell

Welcome to the Eleventh Annual Benjamin and Anne Pinkel Endowed Lecture. The Pinkel Endowed Lecture Series was established through a generous gift from Sheila Pinkel on behalf of the estate of her parents, Benjamin and Anne Pinkel. The Series serves as a memorial tribute to their lives.

Benjamin Pinkel received a Bachelor’s degree in Electrical Engineering, here, from the University of Pennsylvania in 1930.. Throughout his life he was actively interested in the philosophy of the mind and published a monograph in 1992 entitled “Consciousness, Matter, and Energy: The Emergence of Mind in Nature.” The objective of the book was, and I quote, “a re-examination of the mind-body problem in light of new scientific information.” The lecture series is intended to advance the discussion and rigorous study of the deep questions which engaged Dr. Pinkel’s investigations. Over the last ten years, the series has brought in some of the most interesting minds in the field of cognitive science as it pertains to thought, learning, and consciousness. These include Daniel Dennett, Liz Spelke, Martin Nowak, Stan Dehaene, Geoff Hinton, Ray Jackendoff, Colin Camerer, Elissa Newport, and Christof Koch. It’s a great pleasure to add to this list Dr. Alvaro Pascual-Leone who will be speaking about modifying decision-making.

Dr. Pascual-Leone is the Director of the Center for Noninvasive Brain Stimulation and Professor of Neurology at Harvard, the Harvard Medical School, and the Beth Israel Medical Center. He also holds appointments as an Adjunct Professor in Psychiatry and Neurobiology at Boston University and in Cognitive Neuroscience at the Faculty of Arts and Sciences at Harvard. He is also the Associate Director of the Harvard-Thorndike General Clinical Research Center. Dr. Pascual-Leone was born in Valencia, Spain and attended medical school and completed his PhD in Neurophysiology at the University of Valencia and at the Albert Ludwig’s University in Freiburg, Germany. He received his neurobiology training at the University of Minnesota where he also completed a fellowship in clinical neurophysiology. He spent four years at the National Institute for Neurological Disorders and Stroke in Bethesda, Maryland, and joined the Harvard Medical School after working in Spain for four years as an Associate Professor in Physiology at the University of Valencia. As stated on his website, Dr. Pascual-Leone’s overall aim in his research is to understand neural plasticity at a systems level by seeking rules of plasticity that are invariant across neural systems and domains. He has made great strides in doing this via careful and very creative experiments on this topic, such as examining the response of the visual cortex in the blind and even modifications in the response of visual areas in visually unimpaired people who have been temporarily blindfolded. It should be noted that Dr. Pascual-Leone uses a dizzying array of techniques in his research, including PET, MRI, trans-cranial magnetic stimulation, EEG, and MEG, but I hope it will be clear from the talk today, his work, rarely, if ever, loses sight of the forest from all the data trees that these devices generate. In fact, I find many of his papers quite accessible for people who are novices in these technical devices. So, because of this, he has received numerous awards for his work on these topics, but without any further delay let’s welcome Dr. Alvaro Pascual-Leone.

Lecture :: Modifying Decision Making

Alvaro Pascual-Leone

Thank you very much, John, and all of you, and the Pinkel family. It is a pleasure to be here. I hope that what I was planning on telling you will be interesting to you. I am a neurologist from training and so I tend to approach cognitive neuroscience with the ultimate aim of translating it to patient populations, and I think that one talent that we have as a cognitive neuroscience community is actually to try to do that and to try to explore ways by which we can really meaningfully illuminate clinical interventions that oftentimes lag behind the insights that we have from cognitive neuroscience and that could benefit from greater cognitive neuroscience guidance, but, oftentimes, the leap to the clinic fails to be done. So, what I would like to do is start with the problem of the neurobiology in cognitive neuroscience of decision-making and try to make the argument that some of what we are learning can be meaningfully translated to patient populations around the topics of drug abuse and obesity. The idea of how to deal with the fact that we as humans are really bad at making up our mind and making the right decision is by no means unique of neurobiology. It’s broadly in the popular press now, but it dates back as far back as people have started thinking about these issues.

Arguably, one of the first influential positions on what these decision-making challenges is like goes back to Plato, who in Phaedrus describes the chariot allegory: two horses drive us basically in opposite ways—one full of energy and impulse and, oftentimes, making a rush to one direction and another one more thoughtful and calm and making less impulsive decisions—and the challenge is how to control those two forces so that the chariot ultimately moves forward rather than break in pieces. I don’t know about you, but I struggle with that. My sixteen year-old struggles with it even more, and my eleven year-old hasn’t realized yet what many struggles lie ahead. So, the question is what does that ultimately mean? Have we learned something about this? And, of course, many, many disciplines have addressed this, but even social neuroscience now in the conceptualization of dual process models, characterizing a reflexive versus a reflective system, come remarkably close to what resonates in Plato and with different terminologies—a reflexive system that has automatic processes, it’s fast operating, it’s slow in learning, it’s phylogenetically older, it’s linked to brain structures that are largely sub-cortal in [way] and very closely linked to emotional and immediate reward processing, versus a reflective system, or a C-system, that is more controlled processes, slow-operating, it’s faster to learn, therefore, also more plastic, phylogenetically newer linked cortal areas, presumably, particularly engaged in situations of uncertainty, and, perhaps, this is the one horse and the other horse of Plato with new novel characterization.

Now, if you think about it from maybe more anthropological (I am by no means a evolutionary neuroscientist of any sort), but, from a more humanistic point of view, arguably, contrary to the horse allegory, humans are rather unique in that we are able to really control uncertain circumstances (seems to anyway, one of the horses) to the point of going against what would be self-benefit. Altruism, ultimately, is the true selfless giving of one’s self to others for no secondary motives. In that sense that humans are able to do that (well, few are) we become, in some sense, true evolutionary outliers, seems like, that we are driven not by the one horse and the reflexive system motives, but rather are able to control them to a very high degree. So, what does it all mean in terms of brain function, and how can we try to disentangle that? More importantly, perhaps, as a neurologist, if all that is identifiable could we go in and actually modify the amount of control that one system can exert onto the other? Because if we could, aside from illuminating the mechanisms that are at work, that might have meaningful applications to control my sixteen year-old daughter and all the ethics that come with it, but, in addition to that, perhaps to control and help aging populations that have significant problems in that sort of [bias]. Obviously as you start thinking about the possibility of modifying behavior, human behavior, in the decision-making to that degree, it becomes a real ethical issue that you open up as well. I am not going to spend much time in addressing it, but there are substantial issues to discuss as to what constitutes normal behavior, what is appropriate or not appropriate to do, both from an experimental point of view and, ultimately, from an application point of view.

So, what do we know about all this? So, I think that the way we started thinking about it is that, if it is possible to control one form of decision-making—call it impulsive or reflexive for the time being—an overlay on that control that takes into consideration social norms and is more reflective, less impulsive, and fast-learning than, perhaps, the capacity that we’re tapping on, is that of an ability to inhibit more sub-cortical responses. And, of course, there is a whole [like sense] of literature; this is just one summary paper of one aspect of the literature that has argued that there are such cross-modality inhibitory capacities of specific parts of the nervous system, particularly, prefrontal cortex, lateral prefrontal cortex, and, perhaps, particularly, the right lateral prefrontal cortex, across tasks, both in animals, in macaques, and in humans that require a stop-signal or a no-go signal or a task-switching where activities in those areas impose a control that allows for that switch for that no-go to take place. Might it be that such a capacity of this neural substrate is applied not only for specific tasks but across decision-demanding domains? So, in order to address and explore this question, we paired up with Aaron [Sawyer] and Daria Knoch (did their work) and explored the ultimatum game.

The ultimatum game is a neuroeconomics game that some of you may be familiar with. It’s a rather simple game. Two people are brought in and they are going to play with each other. One will be the proposer, the other one the responder. The proposer is given twenty dollars or twenty Swiss Francs and told he or she can offer, or propose, a way to split it. Anything goes. You can say, “I’ll keep twenty and give her nothing,” or you can say, “I’ll split it ten-ten,” or “I’m generous, I’ll give her sixteen and keep four,” or whatever. There are no restrictions. The responder on the other hand can listen to the proposal and can accept it, in which case they’ll both walk out of the room with whatever was agreed upon, or can reject it, in which case, neither one will get anything. That’s the game. It’s very simple. So, I’ll give you twenty and you propose. And most people at the beginning will say twelve-eight; humans are not that generous, we generally don’t split ten-ten. We generally try to get some advantage of it, it seems like. Most of them, in the case of the responder, when they are offered eight and had twelve against them, and so realize it is not quite even, still think “well, better eight than nothing, I’ll take it,” and, so, most people with twelve-eight will accept, even though they will walk out of there thinking, “I’m better off than with nothing, but the guy is not that nice.” However, if the offer is sixteen-four then it gets interesting. How low are you willing to go to walk out of here with something when you came in with nothing, in the face of what you perceive to be unfair? That’s the quirks behind the ultimatum game.

So, basically, you have to balance your self-interest—I want to take off something—against fairness, equity, and reciprocity principles. So, how do we each balance this? Obviously, each one of us will balance it somewhat differently. But as it turns out, from behavioral tasks it is known, that people will reject low offers even if what is at stake is as high as three months their own income, which is quite puzzling. So, maybe in these economic times, it may not be quite as much, but in Switzerland this is what you get. The rejection rates when the offers are below twenty-five percent will be about eighty percent, so most people will reject it—will reject it even though they’re doing something that will harm them.

And so, neuroeconomists have spoken about altruistic punishment. We are as human beings willing to punish another for [what] we believe to be morally wrong even if we are harming ourselves in the process. Sam [Fay] and Jonathan [Cohen] did a study some years ago looking at the brain activity in the responders and asking what brain activity is seen when people go and reject the offer. And what they found was essentially two main areas: the amygdala, the anterior cingulate, and the dorsolateral prefrontal cortex on the other hand, so certain sub-cortical structures of the quote-unquote “reflexive system” and the dorsolateral prefrontal cortex of the reflective system. And they argue that the dorsolateral prefrontal cortex, which they found more strongly activated when the subjects were facing unfair offers, as compared when they were facing fair offers, was there to implement the altruistic punishment. So they say the prefrontal cortex is involved in the control of the emotional impulse to reject the unfair offers. What happens to us is that we get an unfair offer and our blood boils, and if you are a Spaniard particularly, and you say, “no way,” and yet, then, your prefrontal cortex, if you are lucky, will kick in and control you a little bit, give you a little bit of time, then you will still reject it. So, it’s the cognitive control of the emotional impulse that drives you to the [un]fairness goal. That’s how they interpret it.

But, of course, they say a different interpretation is possible, and the different interpretation is that, in fact, the prefrontal cortex is there to inhibit the selfish impulse. So, that, rather than the first thing that comes being, “no way, that’s unfair,” the first thing that comes is, “hey, I get four bucks.” And then you go, “wait, that’s unfair, he’s taking sixteen.” And so, the question is what is the first impulse? Is the first impulse to take, even if it’s unfair, because at least you have something, or is the first impulse to reject because it’s unfair? If the first impulse is the selfish one, then what we need is the capacity to inhibit the selfish impulse under certain circumstances, in a sense, to enable what might be more morally appropriate behavior. So, that’s the question that we were going after. And, of course, consistent with the idea of the right prefrontal cortex implementing some sort of inhibitory control, we hypothesize that this is what goes on, that, as humans, not very different from rats, and certainly not very different from any other animal, the first impulse is to preserve yourself. If you have food at your disposal you’ll reach out for it, if you have four bucks and you had nothing, you’ll reach out for it, and then, as a second elaboration, the cognitive control may kick in. So, how did we go about doing that? Well, if you disrupt the dorsolateral prefrontal cortex in the task, and if Sam [Fay] and Jonathan are correct, then the prediction would be that the disruption will essentially reduce the acceptance of unfair offers, the acceptance rate of unfair offers, because that is the consequence of controlling your rejection impulse. Whereas, if the lateral prefrontal cortex is actually suppressing the selfish impulses, then its disruption will lead to exactly the opposite; it will lead to faster acceptance of more offers, because you basically are not able to reject it anymore.

So, we went about functional MRI, identifying the same areas that Sam [Fay] had identified, then using TMS to target the areas. For the sake of those of you that are not familiar with TMS, that stands for Trans-cranial Magnetic Stimulation. This is Tony Barker who developed the system as we currently use it. In a box like this there’s a band of capacitors. They store an energy of a few thousand amperes and, using an electronic switch, they can be discharged in a matter of milliseconds into a stimulating coil. The shape of the stimulating coil, the geometry of it, determines the geometry of the magnetic field that is generated, and that, in turn, determines the focality of the current that you deliver. If you use an eight-shaped coil like this, and it is appropriately sized and small enough, then you can target an area of about the tip of a little finger, zero point five by zero by five centimeters or so. When you do that and target the motor cortex (we’ll hopefully hear a click), the current passes (there, you heard it), and if you look at his hand, his hand twitched. So, targeting the motor cortex, that current depolarizes the neurons and leads to descending volley contra-lateral that activates the muscles and induces movement. You hopefully see that he is not grimacing in enormous pain; it is well tolerated. And if you are doing it right, it is safe. You do need to follow appropriate guidelines. We want to know exactly where we are on the brain not where we are holding the coil by the subject’s head.

And so, what we do is, we track the position of the subject’s head and the coil with these cameras that are projecting infrared and reflecting it on these little balls. And we digitize the coil and the head, register the subject’s own MRI, and, as you are moving the coil around, it tells you where you’re targeting on the subject’s own brain MRI. You can go around it the other way around and define the region of interest in the MRI, and then, when you are in that region and only then, will it trigger the stimulation. So, you can take the functional MRI of the subject and say, “this is my region of interest, now I’ll move the coil around, when I’m there, it will trigger the stimulation.” We’ve done modeling and I don’t want to bore you with too many physics details, but to give you a little bit of the sense of the issues that are involved (there are a number of technical issues involved), there is a very strong current that you deliver, but you deliver it in a very rapid changing field so that you go from zero to two point five Tesla of magnetic field in a matter of about fifteen milliseconds. And so, that very rapid rate of change determines and induces a current field that is sufficient to depolarize neurons. It’s not the magnetic field that does it; it’s the induced current that does it. The magnetic field is basically just breaching the skin, the skull, and allowing you to get into the brain essentially with very little attenuation, so non-invasively. Depending on the exact tissue characteristics, as Anthony [Watner] has shown in lab, the current distribution is different, and so you need to take into consideration the [Salcack] pattern of each individual, presence of pathology or not, to be sure that you are really targeting the spot you are targeting, but you can actually model that and be certain that you target the appropriate spot. If you do that then there’s things that we have learned as to what happens when you apply trains of stimulation. I’ve shown you what happens when you give one single stimulus, you depolarize the neurons; if it’s in the motor cortex, the hand twitches. If you apply the stimulation repetitively over any one area, depending on your frequency and pattern of stimulation, you leave the cortical area modified. So, you can apply train of stimulations, say targeting this area, apply it at low frequency, one stimulus per second for sixteen hundred stimuli, and then, after you’ve stimulated, the area is rendered depressed in its level of activity, whereas, if you applied in burst of twenty hertz, the area is rendered hyperactive, facilitated after you’ve stimulated. If you look at the effects carefully, like Tony Valero has done, you can see that the effects are not limited to the targeted region, but actually spread from the targeted region across neural networks based on the connectivity between the targeted region and a given area.

So, here, for example, we’re targeting the visual parietal cortex that has a lot of connections to the splenial visual area, and those are excitatory, so you suppress the visual parietal cortex, you withdraw the excitation at a distance, and, therefore, there is a knock-on, trans-synaptic suppression of the visual splenial cortex, whereas neighboring areas that don’t have connections show no impact. So, it is a tool to non-invasively modify activity in a given targeted area, either up or down depending on your stimulation parameters, and, through that targeted area, affect the network, depending on the connectivity. In fact, the distant effect is very telecoupled with the strength of the connections that exist. Now, if you (let me skip through), if you want to know what the connectivity is in any given human individual (this was CAT experiments), you can combine TMS with functional imaging, target a given area with the TMS coil inside the MRI scanner, as done here by [Amir Emidian and John Campodrone], and look at the bold, [vogue] responses to TMS itself, so, you can see here that if you target the visual cortex, you can activate the visual cortex, you can trans-synaptically activate sub-cortical areas (LGN), and you can activate, tie a visual path, dorsal and mental stream, because of this trans-synaptical effects. And, of course, that way you can either study the connectivity or modify it by applying repetitive stimulations to one or multiple areas.

So, let’s go back to the ultimatum game. We’ve identified from the functional imaging study an area in the prefrontal cortex, the same one that is [on the head] shown. We can now use the TMS to guide the stimulation to that area and suppress the activity in that region. So, what happens if you do that versus target the left side? Now, I didn’t point out, but in some face cases, and, in fact, ours, there was activation in both prefrontal cortexes. It’s slightly more on the right than the left, but is bilateral. So, we targeted the left versus the right versus sham stimulation. And you can see that the acceptance rate significantly increases when you suppress the right prefrontal cortex. Not only does the acceptance rate significantly increase, but subjects respond and accept significantly faster when you suppress the right prefrontal cortex. In other words, you tell them, “sixteen-four,” and they go, “yes, I’ll take it!” That’s contradictory to the prediction that this is controlling your emotional response because of the unfairness. Is that sufficient to be sure? Well, we wanted to take it a step further, and, so, in the case of a human proposer, you’re faced with self-interest, on the one hand, for yourself to keep whatever you have, and you have to balance that against the fairness and reciprocity principles related to this other fellow that just came into the lab with you and is doing something that you think is not right. Because, it turns out when playing with this proposer, you just accept it, but you still thought that this was an incredibly unfair thing that the person was doing.

So, what happens if you get the same exact offers but via computer and you are told these are random? So, that was the second behavioral control that we had. Obviously, you’re sifting through, the motives will be the same, you still want to keep the money, but reciprocity really makes no sense, because, although I have to say I often feel like throwing my computer across the room, I know it makes no sense. And fairness and equity, arguably, are also not very important motives, because they don’t carry the day in this case. This is a random generation. So, in that case, the prediction would be that the right prefrontal cortex suppression may have a very different impact, because you are not balancing the two principles, and, indeed, the acceptance rate when you modify the right prefrontal cortex with the TMS versus the left or sham is significantly different. So, when playing with the computer, there is no reciprocity and fairness as opposed to balance against your selfish, if you want to think of it that way. It’s, sort of, impulse intake, and, there, the right prefrontal cortex doesn’t kick in; essentially, it doesn’t come online.

So, what have we learned from this? The implication from this is that, indeed, the right prefrontal cortex is apparently modifying your tendency, your controlling tendency, to impulsively accept what, at least in this task, will lead to a benefit to you—the selfish impulse to take the money. Most of the times, though, these tasks are not played one-on-one; they’re played in a whole group. They’re all interacting with each other, and, in that setting of social context, the behavioral signature of this task has been generally figured out. And that is important because, it turns out, that it matters who exactly you play; it matters what you think about the other person, it matters whether you think that the other person is your boss, not your boss, dresses nicely, doesn’t dress nicely, the gender and all this kind of stuff. So, TMS, to modify activity in any given are (for that matter, functional limiting) to capture the activity, it’s difficult to do in a group with those interactions, but there is another technique of non-invasive brain stimulation that can be used in that setting. So, with Daria, we tried to play the ultimatum game, modifying right versus left prefrontal cortex activity on some of the subjects and, in others, applying sham using trans-cranial direct current stimulation.

So, trans-cranial direct current stimulation goes back, not just to Faraday in the 1800s, but to Galvani and Aldini centuries before. It’s essentially faradizing the cortex. You apply a very small, one to two milliamp current, through an anode and a cathode. You apply it either continuously for some period of time or you put it on and then, after a period of a few seconds, you switch it off. Subjects feel this ramping up of current as an itching in the skin and then they feel nothing else. They feel no difference if you apply anodal stimulation or cathodal stimulation. Most of the current stays in the skin, but enough goes through. There is current in the cortex sufficient to, not depolarize the neurons (it doesn’t cause a twitch, you don’t see anything), but it changes the firing range of these neurons. In fact, it appears to change the membrane potential of the neurons, and so, when another input comes in, if you’ve applied anodal stimulation, there is an increasing firing, if you’ve applied a cathodal stimulation, there is a suppression. So, it’s a pure neural modulatory intervention, and the subjects don’t feel anything; they are truly blunted, and so you can, therefore, play it in this context of applying to different subjects in different ways and seeing how they perform while their prefrontal cortex is modified in its activity.

So, Daria had subjects play under either sham or cathodal TDCS conditions, and what she found was that subjects significantly accepted more offers when the cathodal stimulation was in place, remember, suppressing right prefrontal cortex, than sham. This is four versus sixteen, the same condition as before. The fairness judgment didn’t change; subjects, just like with the TMS, still thought, “this is unfair,” but just took it when the right prefrontal cortex was suppressed in its activity. And that was for each subject when exposed to cathodal stimulation. So, even though the likelihood that each one of us accepts or rejects is variable, any one of us appears to be susceptible to accepting more when our right prefrontal cortex is suppressed. So, from these ultimatum game experiments, we feel that we’ve learned a few things. First, it’s possible to modify the response to this high, complex decision-making processes without altering the fairness judgment that is involved. Suppressing the right prefrontal cortex increases the acceptance, and, so, it is consistent with the hypothesis that the right frontal cortex or the dorsolateral prefrontal cortex suppresses self-centered impulses.

Now, is this as true for the ultimatum game or can it be demonstrated in other tasks? Of course, depending on the task conditions, there is a greater or lesser demand on impulsive, fast decisions, and, so, arguably, it may have to do with the characteristics of the task. How much the dorsolateral prefrontal cortex comes into play rather than the judgment per se? So, we tried to explore this issue with a different task, the Rogers risk task. In this case, subjects are presented with six boxes. The six boxes have different colors, two possible different colors. Each color is associated with a given reward amount that is given to the subjects and explicitly shown to the subjects. And the task is to find the winning token under any one of these boxes, but the probability being identical that the winning token is in any one of these boxes. You are just supposed to give the color. So, in this case, if you say “green,” you have five, six likelihood of being right, but if you are right you will only get twenty bucks. If on the other hand you say “pink,” you’re much less likely to be right, but if you are right, you’ll get eighty. So, subjects know how much risk they are taking, how much reward they can expect from it, and they can take it in consideration when making their decision. So, the level of this can be calculated, the balance of the reward is explicitly told. It’s not really a life situation, most of the time we don’t have that, and we’ll get back into that in a second, but you can then target their same right prefrontal or left prefrontal as a control area, and then have the subjects play for some period of time, and see how modifying activity in the prefrontal cortex changes the performance. What you get is that, when you have suppressed the right prefrontal cortex, subjects make significantly less points than they did with the left or the sham condition. Because of that, they actually end up making significantly less money. The reason is because they make significantly less low-risk choices. They end up risking more, presumably driven by the probability of the reward being significantly higher. So, they see eighty with pink; it’s a much higher risk, chance, but higher potential reward opportunity, and they go with that, even though, because of the way that the task is designed, if you do that, you end up losing. You get less points.

So, suppression of the right prefrontal cortex essentially leads to increased risk-taking behavior in this task, and, because of the design of the task, that is not a good thing. If the task is balanced the other way around, which we have done, then it still leads to impulsive behavior. It still leads to the high-risk behavior being the predominant with suppression of the right prefrontal cortex, but, in that setting, subjects win. So, it’s not about their winning or losing, it’s about the apparent inability to control the tendency to go with the immediate reward and the high risk when they are confronted with the absence of, or in the face of, suppression of the right prefrontal cortex. So that’s nice, but, of course, from the neurological point of view, it would be nicer to do it the other way around. It would be nicer to decrease risk-taking behavior rather than to increase risk-seeking behavior. Can we modify activity in the right prefrontal cortex, increasing it rather than decreasing it, and modify behavior in the opposite direction?

So, Shirley Fecteau has been doing these studies—same task—she’s done it with TMS and with TDCS, these are the results of the TDCS. So, instead of targeting the right prefrontal cortex with the cathode to suppress the activity, she’s applying the anodal stimulation, which will increase the activity. And what happens then is that when the right prefrontal cortex is increasing its activity, subjects make significantly more low-risk choices. And, again, they do that even if you modify the task and make this be a losing proposition. So, it’s not about the outcome; it’s about the way that they go about doing the task. In this particular case, this particular task, the points earned increase, you make more money with the lower-risk choices, but you can play it the other way around, and people will still go with the low-risk choices. So, increasing activity in the right lateral prefrontal cortex decreases risk-taking behavior. So, again, in the risk task, it appears that the lateral prefrontal cortex on the right side is suppressing impulsive self-centered behaviors. But, again, as I mentioned before briefly, this is a situation that is a bit artificial because you know the balance of your risk and, most of the time, we don’t in real life. So, what happens if you do a task where you cannot judge the risk? At least, you cannot be sure of it. The Balloon Analog Task allows you to do that.

This is a task in which the subject sits in front of a computer screen. There is a button and there is a balloon. And you’re told to push the button, and, every time you push, the balloon will pump up, and with every pump, you’ll collect some twenty-five cents, or something like that. At any point you can stop and whatever you’ve collected you will move into permanent storage, but, if you keep going and the balloon bursts, you will lose everything. [Co-edit]. You should know that each balloon has a different bursting point. They are completely unpredictable. You just need to decide whether you’re willing to take the risk. Now, this is a much more ecologically valid thing, because most of the time in real life we don’t know what the risk of the situation really is until it is too late. And this is exactly what happens here, but there are some senses that people develop, and, much like in real life, they believe they actually can gauge how much they can keep going. They really can’t, but they do pretty good. Overall, they pump a number of times, decide to move it into permanent storage, or not, and, if they don’t, they keep pumping again, they will lose it. So, what happens if you modify activity in the right or left prefrontal cortex in this setting? So, in this case, it’s not so lateralized. In fact, the effect is there whether you increase the right, increase the left, or vice versa, it doesn’t really matter; it’s about the balance between left and right that appears to be critical.

But, the net result is that the number of pumps is significantly less when you create an imbalance between the frontal lobes, in the level of activity, by increasing the left and decreasing the right or vice versa. Doesn’t matter which direction you modulate it. So, if you apply a nodal situation to the left and cathodal to the right, or a nodal situation to the right and cathodal to the left, in any case, subjects become more risk averse, as it were, and end up pumping less number of times the balloon and, ultimately, making less money. So, this is the total money earned; it’s significantly less. So, even though they see that they are making less money and they just played the game a number of times before, they still err on the side of making less choices when the stimulation is going on in a particular way. So, the idea is that whether you know the risk or don’t know the risk, whether you can calculate the probability or not, the desire to go with the immediate reward, perhaps the selfish desire, the self-centered behavior, is regulated and controlled, or can be regulated and controlled, by the lateral prefrontal cortex. In some people it is more so, in others it is less so, but, if that is the case, then modifying activity in the right prefrontal cortex can allow us to either exert less or more control over this self-centered behaviors, perhaps, enhancing reflective control over reflexive mechanisms.

Presumably, this is particularly critical in situations where there is an ambiguity conflict, so, decisions, or choice-behaviors, where there are both cultural and moral/social conventions that are opposite to the self-motivated, self-centered behaviors, like the ultimatum game. In that situation, the regulation of the balance between those two, similar to what Plato was talking about, becomes really critical. The argument is that these small experiments that I was showing you reveal a critical role of the right prefrontal cortex in that setting. So, what happens if you have a condition that leads to right dorsolateral prefrontal cortex failure? We have a trauma, frontal dementia, mood disorder, and so forth, where there is evidence that the right prefrontal cortex is dysfunctional, where the presumption would be that you would have impulsive, self-centered behaviors with disregard to social norm and conventions and with disregard to other expectations. If that were the case, then, there is a translational opportunity there. You could use the same type of modulation that I’ve shown you in normal subjects to target the lateral prefrontal cortex and increase the control the id exerts onto sub-cortical areas that, presumably, drives self-centered behaviors.

So, we have tried to model this and proof of principle this with addictive behaviors and eating disorders. Just want to show you some of those results quickly to give you a sense of what the findings look like. So, Philipe Fregni did a study in smokers, initially with acute-provoked nicotine craving and then with, just, spontaneous smoking behavior, tracking their smoking and the acute-provoked nicotine craving, presented a little movie clip of someone smoking and then applied TDCS, the same way that I’ve described, to increase the right dorsolateral prefrontal cortex activity, or left as a control, or sham, and then had the subjects judge again the amount of craving that they experience to the same movies. What he found, is that, over time, from the base line to after the TDCS, in follow-up, there was with the right dorsolateral prefrontal cortex stimulation, but also with the left, a significant small effect, but a significant effect, in decreasing the craving that actually sustained over time. So, not only the right, as we have predicted, but in fact also to the left, increasing activity in right or left dorsolateral prefrontal cortex in each individual subject led to a sustained demonstrable decrease in the craving that was induced by the cue. [Paulo] Poggio has looked at the same type of design for alcohol craving and found essentially the same results, that with the left or the right dorsolateral prefrontal cortex stimulation, increasing activity in either one leads to a significant reduction in the craving compared to sham. John [Compadrone] has looked at it in cocaine addicts, in six cocaine addicts, and found with TMS rather than with TDCS, that the left, in this case, does have the effect, but that the right significantly decreases craving and, in fact, sustains the suppression of the craving for five weeks on average after just five sessions of stimulation. I find the results in obesity particularly interesting, so I want to spend a little more time with that. The argument that Miguel Alonso has put forth is that eating behavior, of course, is critically linked to homeostatic control and reward gratification, but that in humans, in addition, there is a whole cognitive aspect of eating that we often don’t take in consideration in dealing with eating behavior in obesity and that relates to cultural, social messages that we send with food. If you are from Spain and you come to my house, and I put something in front of you to eat, if I don’t, I’m committing a sin, but if I do, and you don’t eat it, you are committing a sin.

And so, therefore, there is a whole elaborate behavior involved with this, which we as humans are able not only to use to our advantage, depending on the situation, but also make it trump all homeostatic control and work gratification principles. The point of killing ourselves by not eating on the basis of, motivated by, political principles—something that seems to go at the very core against nature. If the right prefrontal cortex in all these experiments is playing a role in controlling self-centered behavior, this is the ultimate control mechanism potentially of self-centered behaviors, where you have even a capacity to suppress self-preservation.

In fact, what Miguel has recently done is look at how is food items processed by the brain when you present them in an fMRI context. He had subjects look at a fixation point then presents non-edible or edible items, some are quote-unquote “healthy,” others obesogenic, and they can be presented for a very brief period of time, sixteen milliseconds, or for longer periods of time, two hundred milliseconds, followed by a mask. In a separate session, subsequently to the functional limiting that I want to show you, Miguel has made sure that the subjects, when presented for the sixteen milliseconds, truly don’t see the objects. In fact, don’t have a clue that there are any items presented. (This is shown here.) They are completely random as to whether or not they know what was a picture presented here. The task for the subject is simply to respond to whether these squares, presented subsequently, whether the filled one was on the left or on the right. But what we are interested in, is what is the brain activity related to these items, specifically, to the contrast between these two food items? If you look at brain activity related to the subliminal processing of obesogenic versus non-obesogenic food items, what you get is activity in the right dorsolateral prefrontal cortex. And if you define that region of interest and look at what is correlated with it, you find that any number of risk factors detract personality characteristics that would lead to obesogenicity: eating breakfast away, emotional eating responses, low energy expenditure, family history of obesity, a number of items that are well-recognized in behavioral studies to be predictors of personality, predictors of obesogenity, show a correlation, an inverse correlation, with the amount of activity in the right prefrontal cortex. So, in other words, the less activity the subject shows, the more risk factors they have for obesogenic behavior.

So, if that is the case, what happens when you take any one of us and modify activity in the right prefrontal cortex, just like we were doing, but what happens, is as Philipe and Miguel have done together, that if you do the TDCS, you can basically (just cut to the chase), you can basically show that subjects, when the right prefrontal cortex is increasing its activity by exposure to the anode, but not the other way around, when the left is increased, you get a reduction in food craving when exposed to stimuli of food craving. You can also show that in a spontaneous [unintelligible] of eating, when they go into a buffet and eat as much as they want, they actually eat significantly less. And this is always at the same time, always with the same amount of food during the day. These subjects were admitted to the general clinical research center, we know exactly that they have been eating the same amount of calories. And, finally, which is very interesting too, when you look at the eye movements that the subjects make when exposed to these pictures, they have to explore and answer a given question, to an unrelated question too, turns out, that when they have been exposed, or are being exposed, to the right anode-stimulation enhancing activity in the right prefrontal cortex, they show significantly less time fixating, looking at the food items than at other items.

So, the idea is that the lateral prefrontal cortex appears across tasks, perhaps, to exert repressive control and to self-center behaviors, might be thought of as a switch between reflexive and reflective modes of operation. And it appears that that insight lends itself to translational applications, where we can take those insights and apply them to patient populations that have dysfunctions in the right prefrontal cortex or behavioral dysfunctions related to a lack of control of impulsive behaviors, in such a way that we might offer them a helpful, therapeutic intervention. On the other hand, in broader context, it appears that we’re able to influence and control behavior in humans with this brain stimulation techniques, and, of course, having myself significant trouble controlling how much I eat, the temptation is large to wake up in the morning and do some stimulation before I walk out the door. Jokes aside, it would seem that the temptation to define in some arbitrary way what are desirable behaviors to modify and how to guide decisions, it’s…all these techniques appear to be quite powerful in their ability to modify decisions—they’re non-invasive, they’re relatively simple—and so, as I was saying at the beginning, an ethical debate would seem to be important to open up. Just want to finish up acknowledging again all the people that have actually done all the work and hope that you have lots of questions. Thank you very much.

Questions and Answers

QUESTION 1

First of all, thanks, that was an amazingly interesting talk. And I just had a question about the craving measure and I guess the craving findings. Were you surprised in a way that stimulation of the right prefrontal cortex would actually affect ratings of craving, which I thought, “well, that would be a read-out from the sub-cortical appetitive areas,” right? I mean, unless craving is being interpreted as “would you do it?” collectively, kind of, a revealed preference?

Alvaro Pascual-Leone

Yah, that is a good question. So, we measure it using the visual analog scale, so, basically, it is just how much desire they felt for whatever item—the smoking or the alcohol. What we had predicted is, actually, that the craving would be decreasing. We had predicted the results, but I think it was because we were just simply thinking of the craving as manifestation of the desire to do this, rather then the elaboration of, what you were saying, of how to express the desire I have. I think that more sophisticated measures of craving, obviously, could be done to try to parse out what is really going on. This was just a very preliminary proof of principle, and, similarly, this is very much triggered by the movie and immediately measured. So, I think that it would be interesting to see the craving desire cross situations that expose them to other stimuli and get a sense of what is really modulated.

QUESTION 2

Your headline story is the right side of the prefrontal cortex; it matters. Your results suggest that there is something more nuanced going on. Sometimes, it is different. I’m having a little trouble understanding when the differences matter. Could you help to explain the apparent difficulty?

Alvaro Pascual-Leone

Yah, the subtlety is there. We were hoping for slamdunk results, and it didn’t quite work out that way. So, it is more complex, and, to be honest, you didn’t get it because I don’t know myself. We look at the tasks. I don’t think it is anything specific of the tasks that we could identify. We had not predicted, for example, that with the smoking and with the alcohol, with its bilateral effect, with food only unilateral. You can hand-wave your way around an explanation, which is what I think we did in the papers, which is to say that there is something different about craving that may be more engaging the left side, because some aspects of it are more interpretative of what you are perceiving, and that’s the issue that we are having their separate from the actual desire control that is more right-sided, but that is a tested hypothesis, but it is untested with the results, so I don’t know if that is really the full story.

It’s not only in the application and the translation application that the bilaterally comes out in those two craving studies, where it could be related to the craving. In one of the tasks that Shirley did, the [Bar] task, the effects were also bilateral, and so, it wasn’t only lateral. So, the characteristics of the tasks seem to play a role, and, presumably, that has to mean that there is specific underlying cognitive processes that may be more or less lateralized to one side or the other, and I think that we need to try and disentangle it more. I don’t have a better answer than that. Just perhaps to add one more other thing. There are some data, proof of principal data from other groups, those are targeting only the left side—left versus sham—for smoking, for nicotine, and one for cocaine showing beneficial effects too, so, it may be that, depending exactly on how you set things up and whether or not you look across the two sides, you might have greater bilaterality.

QUESTION 3

Hi, did you find a threshold effect? What is the least amount of stimulation that you could do and still see the effect?

Alvaro Pascual-Leone

Yah, I don’t know the answer to that. So, there are two sides to that question that can be empirically tested. One is you can do a dose-response kind of curve within subject, and which would be very interesting. It’s not a trivial experiment to do, because the residuality of effects is from session to session, and, so, if you do the same subject, you need to wait long enough, otherwise, there will be a build up of the effect of the second session onto the first. But, it can be solved. So, it could be done. The other way to do it is bank on the fact that the effects are individually different, so, even though you stimulate the same parameters, you modulate the amount of cortical activity differently, and so, Miguel is, in fact, now doing a study currently, running the study with obesity, using EEG and FMRI during the TMS, to measure not only the effects that it has behaviorally, but what impact it has genealogically, with the aim of correlating the ultimate impact on behavior with the physiological impact and see whether there is a correlation there.

QUESTION 4

Second one is only slightly frivolous. When you talk about ethics are you also interested in defining non-invasive?

Alvaro Pascual-Leone

Oh, you think that this is invasive?

QUESTION 4

Well, I can imagine that many people could think that that would be invasive, yes.

Alvaro Pascual-Leone

So, the term invasive I use simply from the point of view that we are opening the skin and the skull in targeting the brain, but, obviously, it’s invasive in the sense of invading into your personal sphere in more behavioral ways.

QUESTION 5

[unintelligible]

Alvaro Pascual-Leone

Yah, that’s what I am saying, that in a sense of neurosurgically speaking, it’s non-invasive, even though it clearly is modifying brain activity, we know that, and so, in that sense, it is doing something to the brain. So, there is that ethical aspect of are those effects safe in their own right. I think that there is a lot of data over the last twenty years on non-invasive, in this sense, non-invasive, in general, on brain stimulation to have sense of what, if any, long term effects might be there, and, if you apply the stimulation in a therapeutic aim, multiple sessions in a row without [any effects], there will be lasting impact, but, if you just apply the stimulation one session, short session for a quote-unquote “agnostic” effect, like in the experiments that I was showing you, there are no lasting effects. So, in that sense, I feel comfortable with the safety of it, but, of course, the impact of modifying the behavior of a person may have on that person in a more global sense, it becomes an interesting question, and I don’t know the answer to.

QUESTION 5

Alvaro, throughout the talk, you link the ideas of impulsivity and self-interest together in the sense that a number of your tasks actually, if a person is acting in their short-term self-interest, they’re also acting impulsively. And, I am wondering to what degree you think those emotions are actually linked? Whether you have thought about pursuing experiments that might disassociate those ideas, or, perhaps, manipulating impulsivity in a way that actually works against this notion of self-interest?

Alvaro Pascual-Leone

Thank you, that is a great question. Yah, so I…it didn’t come across properly because I didn’t mean to link them. Some of these tasks actually don’t disentangle them, and so, because of that, the fast response, impulsive response in that sense, ends up being also self-interest driven or maps to the self-interest, but that’s not with all the tasks. So, you can separate some of the tasks, so, for example, the fast response of…it depends on how to define it in the part, you can say, “I’ll take the money now to have a little bit,” it’s more self-interest, whereas they keep on pushing, that is more impulsive, or is it not? So, the speed of response, either one of the two, doesn’t really get modified by the stimulations, so I don’t think it is the impulsivity per se that is being modified by the stimulation, but rather how much self-centered motives are weighed by the individual in performing the task.

In some of the tasks, like the ultimatum game, it turns out that when you are driven by self-impulses of keeping the money, you also respond faster, and so it looks like it is impulsive. Daria has gone on to do some tasks, looking at just the fast versus slow decision-making, where you have to weigh more information, but that the decision is made, or do so quicker. And there is some literature in the faster, more impulsive, not dependant of self motives tasks, there is more activity in the prefrontal cortexes as well in functional imaging, [McClure] and others has shown that. So, in that setting, TMS didn’t affect the decision—it did affect the response time—but it didn’t affect the decision. So, I don’t think that this is a bad impulsivity, to answer your question, but of the content of the decision, but, sometimes, it is not possible to disentangle with these tasks.

QUESTION 6

In the ultimatum game, the reward is measured and it is money, but in the real world, rewards can be complex, intangible, perhaps, like helping a blind person cross a road. How would you account for that?

Alvaro Pascual-Leone

Yah, I think that that is a good question. So, in as much as there is a reward that is translated to a benefit to the individual, I would argue that the same holds true, that this is not just about money, and, hence, the argument that it applies to behaviors, to eating what is in front of you, or taking the pleasure related to the smoke or the alcohol immediately, or whatever. So, in as much as helping somebody cross the street leads to a benefit recognition, being told how nice you are, whatever, I would argue that that would be following similar regulatory controls. So, I would say that the right prefrontal cortex in that setting is still exerting the same control on your drive to do what you think is the thing that will bring you the reward, in this case, helping somebody cross the street.

QUESTION 6

It can also be that people vary in their sense of ethics, so it’s not necessary that the same result may hold for in their case.

Alvaro Pascual-Leone

Yah, there is no question that there is huge individual variability in all these measures and in all these personal characteristics, and I think there are, arguably, I would say, because the amount of right prefrontal control onto sub-cortical structures may be different in different people. So, these measures are all within subject, so, whatever you do, if I activate your right prefrontal cortex more, it seems like you do differently, which is why I’m saying that it opens a whole ethical debate. I think there is great value in the fact that some people are different than others and that we are not a nice homogeneous, everybody’s same amount of, whatever, altruism. So, that’s, in fact, part of the concern, who defines what is appropriate and how do you hold people to it? But, it is within subject that I am talking about, modifying the decisions like that.

QUESTION 6

Thank you.

QUESTION 7

I wonder, if I were to summarize what you’re saying, if I had to tell someone what I heard today, I would say you have done these beautiful studies that show regulating the right or the left prefrontal cortex reduces the tendency to be attracted to a (?) stimulant but, what is going on there I mean why would that be happening? I mean, do you have a theory? Is it like are they thinking more about their future and saying I shouldn’t be doing this now because the future will be bad for me? Why is it that up-regulating, you know, such a gross [intervention] should be so uniformly powerful with the same result?

Alvaro Pascual-Leone

Well, I would turn it around. I would say, I think, we have the same interest in pursuing the question. The question is, to me is given these results, which are surprisingly strong to me. There should be, there is now a whole reason to start trying to disentangle what is really going on. At two levels, I think, both at the biological level, what is the circuitry that actually accounts for this, what are the areas that when you modify the right prefrontal are being impacted? [Mark Halco] is now in the lab now targeting the right prefrontal cortex, or the left, with the TMS and looking at what activation do we see as opposed to what the bolt changes? and how is that change results by engaging the subject in a task like this when you apply the stimulation, it turns out the connectivity, the bolt-induced connectivity is highlighting functional connectivity, is not just anatomy that matters, but the state of the connectivity matters. So if you’re engaging in one task that is going to demand the decision from you down the road using different activities than if the subject is addressed which is not surprising, but may hopefully allow us to find out what is the target area of that modulation? What is the neurobiology of it? At the same time in parallel when you’ve modified the activity in that manner and you decide to go with the different decision, what is driving that decision? What is the cognitive process that the subjects are quite blind to it, so you tell them and their arguments are, well I don’t think this is surprising either, yah, they’ll tell you, “Oh yah, in the ultimatum game, he was a complete jerk, I can’t believe he did that, offer only two or whatever. “ So, why did you accept it? “It just made sense.” But a week ago you didn’t accept any of those offers. “I didn’t feel like it then, but it was the right thing to do now. “ OK. There is no link established even to the stimulation. A few subjects, when you go through this questioning session at the end they will say well I presume it must have been whatever you were doing. But there is no insight whatsoever. So basically we need to design tasks that will capture the phenomena with other tasks that presumably try to characterize also what else is going on. We’ve tried to do some things, so for example we know that the go, no-go is not affective, is not simply an across-the-board inhibition regulation. That’s not it. It’s more complex than that, at least for the same amount of stimulation that we’re applying. That’s a good question.

John Trueswell

If there are no more questions, let’s thank Alvaro again. And if you have other questions that you want to talk to him one-on-one about, there’s a reception just outside here.

 

University of Pennsylvania