Funded Grants

Human Emotion at the Level of the Single Neuron

Imagine for a moment that you have just received a letter informing you of a large award from the McDonnell Foundation to support your research. They are giving you hundreds of thousands of dollars that you desperately need to do the experiments you've always dreamed of doing. How do you feel? What are you thinking? What might you decide to do next? It seems apparent that your thoughts and your behavior would be greatly influenced by the emotion that you feel upon hearing this news. No doubt, you would have felt, thought, and acted differently had the news been negative.

One doesn't have to win an award or to lose a loved one to realize that nearly every waking moment of our lives is imbued with emotion. While we can all agree that emotions influence our lives, it is somewhat unclear what exactly they are good for. Does it help you to feel elated at winning the award, or would life be just the same if we had no emotions?

A fundamental problem that all organisms face is how to select, from a plethora of stimuli as inputs and a wide array of possible behaviors as outputs, which actions to take in response to which stimuli. Without such an ability to assign relevance to stimuli and actions, any system is quickly overwhelmed--a problem that confronts those who are attempting to construct artificially intelligent systems today. The brain has a number of ways of providing selection; that selection which occurs on the basis of the value (ultimately, value in terms of an organism's survival) of a particular stimulus or action is the core topic of emotion. Emotion guides our attention towards what is important, permits us to remember what matters, and leads us to make choices that, more often than not, are in our best interest. A human life entirely without emotion would be not just bland, it would mean death.

In our above example, it is clear that multiple events would have occurred: you perceived the words in the letter, your brain processed their meaning and triggered an emotion, and the emotion altered your attention, your memory for the event, and the decisions you subsequently made. From introspection, as well as from studies in cognitive psychology, we know something about the component processes described here. The question I want to address is how these processes are implemented in the human brain.

Answers to this question have been slow to emerge, in large part a consequence of the relative neglect of emotion within cognitive neuroscience, until rather recently. Ever since Aristotle described humans as rational animals, emotion has been viewed as distinct from reason. In the past decade or two, we have begun to realize that emotion is simply one particular aspect of how the brain processes information, an aspect that permits an organism to link representations of external stimuli and events with self-representations of its body state. It is this juxtaposition of information about the outside world and about the internal world of the organism's own body that permits an evaluation of stimuli in terms of the survival-related meaning they have for the organism.

Recent findings in the cognitive neuroscience of emotion have come from studies of patients with damage to particular regions of their brain, as well as from functional imaging studies (such as PET and functional magnetic resonance imaging) of normal individuals. These studies have indicated that there are certain structures in the brain, notably including the amygdala and the orbitofrontal cortex, that are responsible for associating sensory stimuli (like the words on the page) with their emotional meaning. Patients with lesions to these regions of the brain may be unable to recognize emotions from stimuli, and may be unable to guide their behavior on the basis of how they feel; consequently they are often severely impaired in their everyday lives.

While prior studies of human emotion have pointed to particular structures, they have been unable to describe how emotion is processed in great detail. In particular, they have been unable to investigate precisely the detailed mechanisms whereby the cells within these structures process emotional information at various points in time. These are questions which have instead been addressed in animals, by inserting electrodes into the brain to measure the responses of individual neurons. In the past four years, my colleagues and I have begun to apply microelectrode recordings from single neurons also to the human brain.

Such studies, of course, cannot be done solely for research purposes, but rely on a clinical setting. All of our patients have epilepsy that cannot be controlled adequately by medication, and are thus candidates for neurosurgical resection of the brain tissue that is causing the epilepsy. To guide such a resection, the neurosurgeon implants electrodes into their brains so that the location of seizure foci can be accurately determined. However, the patients can also participate in our research studies at the same time that they undergo clinical monitoring. The electrodes are typically in the brain for 1-2 weeks, and during that time we show the patients pictures, have them listen to sounds, and have them watch film clips while we record from neurons in their brain.

Our findings so far have been quite dramatic. In one series of studies, we showed patients pictures that varied in terms of their emotional content--pleasant pictures of babies and food, neutral pictures of chairs and tables, and aversive pictures of dead or mutilated people. When we recorded from neurons in the orbitofrontal cortex, we observed a change in the firing pattern of neurons only when the patient saw the aversive pictures, but not when he saw pleasant or neutral pictures.

Several aspects of this experiment were novel. First, we were able to localize the precise region of the brain from which we recorded these responses. Our electrodes had electrical contacts at particular locations, and we obtained magnetic resonance scans of the patient's brain after the electrodes had been implanted. Second, we measured the response from a neuron with millisecond precision. In this way, we were able to show that the responses to the aversive pictures first occurred 120 milliseconds after the picture appeared on the screen. This is amazingly fast: it is almost certainly faster than it would take the patient to become consciously aware of the picture that he saw. In fact, 120 milliseconds is so fast that the brain only has time for a handful of processing steps--yet, evidently, that small number of steps is sufficient to permit the brain to discriminate the picture as aversive.

The issue of timing is of very general importance. Think back to our original example of the letter. What happens in time? Clearly, you don't start out feeling joyous and relieved (if you are like me, when you first open the letter you feel rather tense and worried). Then you see the words in the letter. A second later you do indeed feel joyous and relieved, and this emotion evolves over the course of the next several seconds or minutes and then gradually subsides. What events in your brain correspond to these changes? We are investigating this question by recording the activity from several brain structures over some length of time after the initial appearance of a stimulus. Our eventual goal is to describe the flow of information in the brain: where in the brain and at what point in time is information about the perceptual visual attributes of a stimulus processed? Where and when is information about the emotional significance of the stimulus processed? And what are the neuronal correlates of subsequently encoding the stimulus into memory or making a particular decision to act? While these are large and difficult questions, to be sure, they are becoming tractable as we combine methods such as studies of lesion patients, or studies involving functional imaging, with rare opportunities to record from single neurons in neurosurgical patients.

In addition to answering questions about how emotion is processed by healthy brains, our research has important implications for disorders in which emotion is processed pathologically. Nearly every neuropsychiatric disorder features a dysfunction in emotion: schizophrenia, autism, depression, phobias, and post-traumatic stress disorder are all, to a large extent, disorders of emotion. How is it that some people are unaffected by certain situations or stimuli, whereas others are plunged into depression or panic? We believe that a part of the answer will involve pathological function within the prefrontal cortex and the amygdala, structures whose detailed function our studies are beginning to elucidate.

But the overall aims of research on emotion go beyond our curiosity about how the brain works, and go beyond accruing information that can be of clinical help. Emotions are so pervasive, and play such a key role in activities ranging from politics to science to art, that a better understanding of emotion in a very general sense will give us a deeper understanding of who we are as human beings.