Computational Neurophysiology Laboratory

Christopher D. Fiorillo Assistant Professor Department of Bio and Brain Engineering Korea Advanced Institute of Science and Technology 335 Gwahangno Yuseong-gu Daejeon 305-701 KOREA Tel:+82-42-350-4326 Fax: +82-42-869-8680 fiorillo@kaist.ac.kr

OPEN POSITIONS

I am currently seeking interns, graduate students, and post-doctoral fellows, for work in each of the areas described below.

RESEARCH

The computational principles underlying neural function are largely unknown. Work in our laboratory combines computational principles with neurophysiological and behavioural analyses to further our understanding of neural information processing. The purpose of the brain is to select amongst potential behaviours or motor outputs, a function that could be viewed as “decision-making” (in a broad sense of the term). The basic problem in decision-making is uncertainty about aspects of the world related to biological goals, or “reward.” Thus our research seeks to understand how neurons and networks of neurons can learn through neural plasticity to accurately predict reward-related aspects of the world.

Testing a General Computational Theory of the Nervous System

A primary goal is to test a general computational theory of nervous system function. My previously published work (Fiorillo, 2008), which is freely available at www,plosone.org, proposed a general computational theory of the nervous system. According to the theory, each neuron learns in the same manner (through Hebbian and anti-Hebbian plasticity) to predict the state of a small part of the world. However, because each neuron develops in a unique environment, each neuron naturally acquires its own distinct information about a distinct part of the world, depending on the statistical patterns in the inputs to which the neuron is exposed. Due to the influence of reward feedback on synaptic plasticity, neurons further from the sensory periphery would have less information about the immediate sensory world, but more information about “future reward,” and would thus be in a position to render the system’s “decision” about the most appropriate output. If the theory is correct, it could potentially allow us to progress towards the creation of artificial neural networks that possess the intelligence of biological nervous systems.

One critical means of testing the theory will be to simulate a network of these artificial neurons and to ask whether the network is able to organize itself so as to generate intelligent and adaptive outputs. The theory also proposes a novel and important role for non-synaptic ion channels. Past theoretical and experimental work has focused on plasticity at synapses as a critical component of learning. Different synapses contribute information from different points in space, and an individual neuron selects some synapses over others. Similarly, different types of voltage-regulated non-synaptic ion channels contribute information from different periods of the past (due to their differing kinetic properties), and somehow a neuron selectively expresses some types of ion channels but not others. However, although the effect of non-synaptic ion channels on a neuron’s output is comparable in strength to the effect of synaptic ion channels, little attention has been given to the rules that govern how a neuron selects from amongst its non-synaptic ion channels. My published theory suggests that the detection of coincident activation of non-synaptic ion channels and the neuron functions to select those ion channels that provide the most predictive temporal information (following a plasticity rule known as “anti-Hebbian”). This novel proposal will be tested through in vitro electrophysiological experiments.

Dopamine Neurons

In addition to exploring the general computational theory described above, a second component of the work in our laboratory examines the physiology of midbrain dopamine neurons in behaving animals. According to the theory summarized above, as well as other accounts, the neural development of goal-directed behavior requires one or more reward signals that shape neural circuitry. A group of neurons in the midbrain that contain the neurotransmitter dopamine are thought to provide such a reward signal. Dopamine neurons are well known to be of primary importance in drug addiction, Parkinson’s disease and schizophrenia. Starting around 1980, dopamine became known to the public as the “pleasure chemical,” although we now know that this description is rather simplistic and misleading. Physiological studies have shown that dopamine neurons are activated when reward value is better than expected, and their activity is suppressed when reward value is worse then expected. Thus dopamine neurons are said to encode a “reward prediction error.” Prior to this physiological discovery, such errors were already used to drive learning in models of reinforcement learning, in both machines and animals. Based on the apparent correspondence between theory and physiology, as well as a large body of pharmacological evidence, it is believed that dopamine may function to teach the brain to distinguish what is “good” from what is “bad” by modulating neural plasticity.

To investigate the function of dopamine neurons, electrophysiological recordings of individual dopamine neurons are performed in behaving animals. My published work has examined the effect on dopamine neurons of uncertainty about reward magnitude (Fiorillo et al., 2003; Tobler et al., 2005), and has used the dopamine error signal to characterize the temporal precision of reward prediction (Fiorillo et al., 2008) (see figure). My unpublished work has shown that stimulation of a structure in a part of the brain called the thalamus has reward value and also activates midbrain dopamine neurons (Fiorillo and Newsome, 2006). Although the latency and amplitude of this activation was similar to the activation caused by natural rewards, the activation was not diminished when electrical stimulation was predictable, unlike the case of activation to natural rewards. Preliminary results from another study show that the activity of dopamine neurons tends to be suppressed by aversive stimuli (Fiorillo and Newsome, 2008). My observations suggest that dopamine neurons are sensitive to a variety of different types of reward events, and I have characterized how an animal’s prior information related to uncertainty about reward magnitude and timing shapes the prediction-error responses of dopamine neurons.

Whereas past work has examined responses of dopamine neurons to natural reward stimuli, an important component of future work will be to study the influence of artificial reward stimuli on dopamine neurons, including electrical brain stimulation reward and addictive drugs. An important goal of this work is to explore computational theories of reinforcement and drug addiction. Another research interest is in the effect of various clinically relevant drugs on the responses of dopamine neurons to natural reward stimuli. In addition, future experiments will examine the effect of the dopamine released by natural reward events on behavior and neural plasticity.

References

Fiorillo CD, Newsome WT. Modulation of dopamine neurons by aversive stimuli. 38^th annual meeting of the Society for Neuroscience, Washington D.C., Nov. 15–19 (2008).

Fiorillo CD. Towards a general theory of neural computation based on prediction by single neurons. PLoS ONE, 3: e3298 (2008).

Fiorillo CD, Newsome WT, and Schultz W. The temporal precision of reward prediction in dopamine neurons. Nature Neurosci 11: 966-973 (2008).

Fiorillo CD, Newsome WT. Activation of Midbrain Dopamine Neurons by Brain Stimulation Reward in Mediodorsal Thalamus. 36^th annual meeting of the Society for Neuroscience, Atlanta, Oct. 14-18 (2006).

Tobler PN, Fiorillo CD and Schultz W. Adaptive coding of reward value by dopamine neurons. Science 307: 1642 – 1645 (2005).

Fiorillo CD, Tobler PN, and Schultz W. Discrete coding of reward probability and uncertainty by dopamine neurons. Science 299: 1898-1902 (2003).

Curriculum Vitae

March 10, 2009

Name:	Christopher Dante Fiorillo
Position:	Research Associate
Date of Birth:	January 28, 1971
Place of Birth:	Eugene, Oregon, U.S.A.

Education:

Institution	Degree	Year	Field
Oberlin College	B.S.	1993	Neuroscience
Oregon Health Sci. Univ.	Ph.D.	2000	Neuroscience

Research Experience:

Position	Mentor / Dept.	Institution	Dates
Assistant Professor	Bio Brain Engin	K.A.I.S.T., South Korea	2009-present
Research Associate	W. T. Newsome	Stanford University, CA	2004-2008
Research Associate	Wolfram Schultz	Cambridge University, England	2002-2003
Postdoctoral Fellow	Wolfram Schultz	University of Fribourg, Switzerland	2000-2001
Graduate Student	John T. Williams	Vollum Inst, Oregon Health Sci Univ	1994-1999
Lab Rotations	various	Oregon Health Sciences Univ	1993-1994
Summer Fellow	Richard Baird	RS Dow Neurological Sci Inst, OR	1993
Undergraduate Student	Denny Smith	Oberlin College, Oberlin, OH	1992 –1993

Publications:

Fiorillo CD. Dopamine neurons and the biophysical basis of temporal prediction. In Attention and Time, Nobre and Coull, eds. Oxford, England: Oxford University Press (in press).

Schultz W, Preuschoff K, Camerer C, Hsu M, Fiorillo CD, Tobler PN, and Bossaerts P. Explicit neural signals reflecting reward uncertainty. Phil Trans R Soc B 363: 3801-3811 (2008).

Fiorillo CD. Towards a general theory of neural computation based on prediction by single neurons. PLoS ONE, 3: e3298 (2008).

Fiorillo CD, Newsome WT, and Schultz W. The temporal precision of reward prediction in dopamine neurons. Nature Neurosci 11: 966-973 (2008).

Fiorillo CD, Tobler PN, and Schultz W. Evidence that the delay-period activity of dopamine neurons corresponds to reward uncertainty rather than backpropagating TD errors. Behavioral and Brain Functions 1: 7 (2005).

Tobler PN, Fiorillo CD and Schultz W. Adaptive coding of reward value by dopamine neurons. Science 307: 1642 – 1645 (2005).

Fiorillo CD The uncertain nature of dopamine. Molecular Psychiatry 9: 122 - 123 (2004).

Fiorillo CD, Tobler PN, and Schultz W. Discrete coding of reward probability and uncertainty by dopamine neurons. Science 299: 1898-1902 (2003).

Paladini CA*, Fiorillo CD*, Morikawa H and Williams JT. Amphetamine blocks glutamate inhibition of midbrain dopamine neurons. Nature Neurosci 4: 275-281 (2001). (*equal contributions)

Fiorillo CD and Williams JT. Cholinergic inhibition of ventral midbrain dopamine neurons. J Neurosci 20: 7855-7860 (2000).

Fiorillo CD and Williams JT. Selective inhibition by adenosine of mGluR IPSPs in dopamine neurons after cocaine treatment. J Neurophysiol 83: 1307-1314 (2000).

Fiorillo CD and Williams JT. Glutamate mediates an inhibitory postsynaptic potential in dopamine neurons. Nature 394: 78-82 (1998).

Fiorillo CD, Williams JT, and Bonci A. D1 receptor regulation of synaptic potentials in the ventral tegmental area after chronic drug treatment. Adv. Pharmacol. 42: 1002-1005 (1998).

Fiorillo CD and Williams JT. Opioid desensitization: interactions among G-protein-coupled receptors in the locus coeruleus. J Neurosci 16: 1479-1485 (1996).

Public Presentations of Unpublished Work:

Fiorillo CD, Newsome WT. Modulation of dopamine neurons by aversive stimuli. 38^th annual meeting of the Society for Neuroscience, Washington D.C., Nov. 15–19, 2008.

Fiorillo CD, Newsome WT. Activation of Midbrain Dopamine Neurons by Brain Stimulation Reward in Mediodorsal Thalamus. 36^th annual meeting of the Society for Neuroscience, Atlanta, Oct. 14-18, 2006.

Thesis:

The synaptic regulation of ventral midbrain dopamine neurons and its modulation by repeated cocaine treatment. Oregon Health Sciences University, 2000.

Grants and Awards:

“World Class University” Project grant from the Korea Science and Engineering Foundation with Doheon Lee, Kwang-Hyun Cho, and Yong Jeong (2009-2013) (Grant code: R32-2008-000-10218-0).

Human Frontiers Science Program Long-Term Fellowship (2000 – 2003)

National Research Service Award, National Institute on Drug Abuse (2000)

(individual postdoctoral, never activated)

Paper of the Year Award, Oregon Health Sciences University, 1998

National Research Service Award, National Institute on Drug Abuse (1997 – 1999)

(individual predoctoral)

Neuroscience Scholar, Oregon Health Sciences University, (1993 – 1994)

Coleman Summer Fellowship, R.S. Dow Neurological Sciences Institute, Portland, OR. (1993)

Patent:

Fiorillo CD, “Prediction by single neurons and networks” U.S. Patent Application 12/271,282, November 14, 2008.

Fiorillo CD, “Prediction by single neurons and networks” Patent Cooperation Treaty Application, November 20, 2008.