Our research in this area spans a wide spectrum, from Bayesian methods and theories of sensory-motor learning and control to neural networks, information encoding and decoding, and biophysical modeling of cellular electrophysiology. Some of our faculty in this area are also involved in brain-machine interfaces and systems neuroscience research.
Investigating the mechanisms of motor output the spinal cord in both normal and disease states
Neurons in the spinal cord provide the neural interface for sensation and movement. Our lab focuses on the mechanisms of motor output in both normal and disease states (spinal injury, amyotrophic lateral sclerosis). We use a broad range of techniques including intracellular recordings, array recordings of firing patterns, 2-photon imaging, pharmacological manipulations, and behavioral testing. These techniques are applied in in vitro and in vivo animal preparations. In addition we have extensive collaborations with colleagues who study motor output in human subjects.
For lab information and more, see Dr. Heckman's faculty profile.
See Dr. Heckhman's publications on PubMed.
Contact Dr. Heckman at 312-503-2164.
Interested in how the brain controls movement, motor learning, and problem solving activities
The brain has a remarkable capacity for learning and controlling complex movements and thought, using current events and memories of past experiences. To do this, it uses neuroanatomically modular loops called Distributed Processing Modules (DPMs) that link different regions of the cerebral cortex to specific loops through the basal ganglia and the cerebellum (Figure 1). This diagram of DPM architecture schematically summarizes the known neuroanatomy as an array of DPMs (cf. Houk 2005 for additional details).
Figure 1: The DPM Architecture of the brain
The bi-directional green arrows represent the predominantly excitatory reciprocal connections between related areas of the cerebral cortex. Cortical area M1 is highlighted in this diagram because it is the main source of voluntary motor commands for limb movements, and because of this we understand the neural operations that contribute to limb movement learning and control particularly well.
Four of the many loops between different areas of cerebral cortex and the basal ganglia are illustrated by the bi-directional red arrows. Red is used to signify that these loops have multiple inhibitory stages, specialized for action selection through disinhibition and for action de-selection through inhibition of disinhibition. This circuit is complicated, but behaviorally it is critical for insuring that we do not try to do too many things at once.
Most areas of cerebral cortex also have loops through the cerebellum, which are illustrated by bi-directional blue arrows. Blue is used to signify that each loop is actually the combination of two loops, one excitatory through the cerebellar nucleus that is specialized for nonlinear amplification, and the other inhibitory through the cerebellar cortex that is specialized for refinement. These special operations of specific stages of the loops through basal ganglia (BG) and cerebellum (CB) are explained more fully in Figure 2, which represents any one of the many DPMs in the brain.
Figure 2: Learning and control operations in a single DPM
Hebbian learning occurs in the cerebral cortex where the control operation is pattern formation. Reinforcement learning occurs in the basal ganglia (BG). Furthermore, the main BG control operation is pattern classification, which occurs mainly in the striatum based on cortical and thalamic input to its spiny projection neurons. Through direct pathways (disinhibition) and indirect pathways (inhibition of disinhibition), a coarse selection of goals is discovered as a consequence of dopamine neuromodulation (purple diamond signifying reward prediction). These representations of goal discovery are briefly stored in reciprocal corticothalamic pathways while being sent to the cerebellum (CB) to generate an intention. Supervised learning occurs in the refinement stage in CB cortex through long-term depression of parallel fiber / Purkinje cell synapses (purple diamond representing error correction). The positive feedback loop between the cerebellar nucleus and the cerebral cortex is essentially a bistable working memory of potential goals that is refined by prominent inhibition from Purkinje cells in the cerebellar cortex.
In summary, BG loops discover opportune goals through reinforcement learning, and CB loops generate intentions capable of achieving these goals through supervised learning. These operations are modular and apply to most areas of the cerebral cortex. Thus, the intentions that are sent as output from the module can be motor commands, motor plans, working memories, or other contributions to problem solving activities.
Our DPM model of brain function is founded on:
- the neuroanatomy of brain pathways and their neurons
- rules of synaptic plasticity, cellular / molecular neurophysiology, and the biophysics of neurons
- recordings of the messages transmitted along specific sensory-motor pathways in behaving animals using microelectrodes
- functional imaging showing task-specific activity of brain networks in human subjects
Our goal is to build a coherent theory of motor learning and control, and to extend our findings to cognitive neuroscience and problem solving, taking advantage of the analogies based on the anatomical similarity of the neural circuits.
For additional career information, see James C Houk, PhD, faculty profile.
- Caligiore D, Pezzulo G, Baldassarre G, Bostan AC, Strick PL, Doya K, Helmich RC, Dirkx M, Houk J, Jorntell H, Lago-Rodriguez A, Galea JM, Miall RC, Popa T, Kishore A, Verschure PF, Zucca R, and Herreros I. (2017). Consensus paper: Towards a systems-level view of cerebellar function: The interplay between cerebellum, basal ganglia, and cortex. Cerebellum 16: 203-229.
- Schwab DJ and Houk JC (2015). Presynaptic inhibition in the striatum of the basal ganglia improves pattern classification and thus promotes superior goal selection. Front Syst Neurosci 9: 152
- Houk JC (2012) Action selection and refinement in subcortial loops through basal ganglia and cerebellum. In: Modelling natural action selection (chapter 10), edited by Seth AK, Prescott TJ, and Bryson JJ, Cambridge University Press, Cambridge, pp. 176-207.
- Scheidt RA, Zimbelman JL, Salowitz NM, Suminski AJ, Mosier KM, Houk J, and Simo L (2012) Remembering forward: Neural correlates of memory and prediction in human motor adaptation. Neuroimage 59: 582-600.
- Keifer J and Houk JC (2011) Modeling signal transduction in classical conditioning with network motifs. Front. Mol. Neurosci. 4:9. doi: 10.3389/fnmol.2011.00009
- Hill, S. K., B. A. Griffin, J. C. Houk and J. A. Sweeney (2011). "Differential effects of paced and unpaced responding on delayed serial order recall in schizophrenia." Schizophrenia Research 131: 192-197.
- Houk, J. C. (2011). "Syntax in the brain: Motor syntax agents." Proceedings of the Eighth International Conference on Complex Systems NECSI: 1462-1476.
- Fraser, D. and J. C. Houk (2011). "Motor syntax disorder in schizophrenia." Proceedings of the Eighth International Conference on Complex Systems NECSI: 1516-1529.
- Ohta, H., Y. Nishida and J. C. Houk (2011). "Presynaptic inhibition and incremental learning in the striatum of the basal ganglia." Proceedings of the Eighth International Conference on Complex Systems NECSI: 1509-1515.
- Houk, J. C. (2011). "Can DPM brain agents write stories and sing songs?" Proceedings of the Eighth International Conference on Complex Systems NECSI: 1539-1548.
- Houk JC (2010). Voluntary Movement: Control, Learning and Memory. Encyclopedia of Behavioral Neuroscience. G. F. Koob, M. Le Moal and R. F. Thompson. Oxford, Academic Press. 3: 455-458.
- Botvinick M, Wang J, Cowan E, Roy S, Bastianen C, Patrick Mayo J, Houk JC (2009). An analysis of immediate serial recall performance in a macaque, , Animal Cognition 12:671-678
- Tunik E, Houk JC, Grafton ST. (2009). Basal Ganglia Contribution to the Initiation of Corrective Submovements.NeuroImage, 47: 1757-1766
- Wang J , Dam G, Yildirim S, Rand W, Wilensky U, Houk JC (2008). Reciprocity Between the Cerebellum and the Cerebral Cortex: Nonlinear Dynamics in Microscopic Modules for Generating Voluntary Motor Commands.Complexity 14(2): 29-45.
- Houk JC, Bastianen C, Fansler D, Fishbach A, Fraser D, Reber PJ, Roy SA, Simo LS (2007). Action selection in subcortical loops through the basal ganglia and cerebellum. Phil. Trans. R. Soc. B 362: 1573-1583.
- Houk JC (2007) Models of Basal Ganglia. Scholarpedia, 2(10):1633
- Houk JC (2007) Biological Implementation of the Temporal Difference Algorithm for Reinforcement Learning: Theoretical Comment on O’Reilly et al. Behavioral Neuroscience Vol. 121, No. 1, 231–232.
- Fishbach A, Roy SA, Bastianen C, Miller LE, Houk JC. (2007) Deciding when and how to correct a movement: discrete submovements as a decision making process. Exp. Brain Res.177:45-63
- Houk JC. (2005) Agents of the Mind. Biol. Cybern. 92: 427-437.
- Holdefer RN, Miller LE, Houk JC. (2005) Movement-Related Discharge in the Cerebellar Nuclei Persists After Local Injections of GABAA Antagonists. J. Neurophysiol 93:35-43.
- Fraser D, Park S, Clark G, Yohanna D, Houk JC. (2004) Spatial serial order processing in schizophrenia. Schizophrenia Research. 70:203-213.
- Houk JC, Mugnaini E. (2003) Cerebellum. In Larry Squire's Fundamental Neuroscience, V. Motor Systems, Chapter 32. Elsevier Science, pp.1-46.
- Novak KE, Miller LE, Houk JC. (2002) The use of overlapping submovements in the control of rapid hand movements. Exp Brain Res 144:351–364.
- Houk JC, Miller LE. (2001) Cerebellum: Movement Regulation and Cognitive Functions. In: Encyclopedia of Life Sciences.
- James C. Houk, Andrew H. Fagg, Andrew G. Barto (2000) Fractional Power Damping Model of Joint Motion.
- Sherwin E. Hua, James C. Houk, Ferdinando A. Mussa-Ivaldi (1999) Emergence of symmetric, modular, and reciprocal connections in recurrent networks with Hebbian learning. Biol. Cybern. 81, 211-225
- Beiser DG, Houk JC. (1998) Model of cortical-basal ganglionic processing: encoding the serial order of sensory events. J Neurophysiol 79:3168-3188.
- Hua SE, Houk JC. (1997) Cerebellar guidance of premotor network development and sensorimotor learning.Learn.Mem. 4: 63-76.
- Houk JC,Buckingham JT, Barto AG. (1996) Models of the cerebellum and motor learning. Behavioral and Brain Sciences 19, 368-383.
- Houk JC, Alford S (1996) Computational Significance of the Cellular Mechanisms for Synaptic Plasticity in Purkinje Cells. In: Behavioral and Brain Sciences. 19(3): 457-461.
- Houk, JC, Adams, JL, Barto, AG. (1995) A Model of How the Basal Ganglia Generate and Use Neural Signals that Predict Reinforcement. In Models of Information Processing in the Basal Ganglia. JC Houk, JL Davis, DG Beiser, eds., The MIT Press, pp. 249-270.
- Houk JC, Wise SP. (1995) Distributed modular architecture linking basal ganglia, cerebellum and cerebral cortex: Its role in Planning and controlling action. Cerebral Cortex 5: 95-110.
- James C. Houk, Joyce Keifer and Andrew G. Barto (1993) Distributed motor commands in the limb premotor network. Trends in Neurosciences Vol. 16: pp27-33.
- Houk, JC (1991) Outline for a theory of motor learning. Tutorials in motor neuroscience, edited by GE Stelmach, and J Requin. The Netherlands: Kluwer Acad. Publ., pp. 253-268.
- Houk, JC (1989) Cooperative control of limb movements by the motor cortex, brainstem and cerebellum. Models of brain function. RMJ Cotterill. Cambridge Univ Press, pp. 309-325.
- Houk JC (1988) Control strategies in physiological systems. FJ Reviews 97-111.
- Houk JC, Rymer, WZ (1981) Neural Control of Muscle Length and Tension. Handbook of Physiology--The Nervous System II. V.B. Brooks. Bethesda, MD, Am. Physiol. Soc.: 257-323.
- Houk JC (1979) Regulation of Stiffness by Skeletomotor Reflexes. Annual Reviews Journal. 99- 114.
- Houk JC (1978) Participation of Reflex Mechanisms and Reaction Time Processes in the Compensatory Adjustments to Mechanical Disturbances. Cerebral Motor Control in Man: Long Loop Mechanisms, Prog.clin. neurophysiol, vol 4. 193-215.
Understanding the nature of the somatosensory and motor signals within the brain that are used to control arm movements
The primary goal of the research in my lab is to understand the nature of the somatosensory and motor signals within the brain that are used to control arm movements. Most of the experiments in my laboratory rely on multi-electrode arrays that are surgically implanted in the brains of monkeys. These “neural interfaces” allow us to record simultaneously from roughly 100 individual neurons in the somatosensory and motor cortices and thereby study the brain’s own control signals as the monkey makes reaching and grasping movements. We can also pass tiny electrical currents through the electrodes to manipulate the natural neural activity and study their effect on neural activity and the monkey’s behavior.
Current projects seek to understand:
- How motor cortical activity leads to the complex patterns of muscle contractions needed to produce movement
- How movement of the limb and forces exerted by the hand are “encoded” in the activity of neurons in the somatosensory cortex
We also study how these relations are affected by behavioral context: the magnitude and dynamics of exerted forces, the varied requirements for sensory discrimination, and the quality of the visual feedback that is provided to the monkey to guide its movements.
Along with this basic research, we can use these neural interfaces to bypass the peripheral nervous system, in order to connect the monkey’s brain directly to the outside world. We are developing neural interfaces that ultimately will use signals recorded from the brain to allow patients who have lost a limb to operate a prosthetic limb. The interface may also be used to bypass a patient’s injured spinal cord in order to restore voluntary control of their paralyzed muscles. Conversely, electrical stimulation of the brain will restore the sense of touch and limb movement to patients with limb amputation or spinal cord injury. This highly interdisciplinary work is enabled by numerous collaborations at Northwestern University and other institutions.
See. Dr. Miller's publications on PubMed.
Contact Dr. Miller at 312-503-8677.
Investigating the sensory-motor system through a close interaction with artificial systems
Our laboratory (the Robotics Lab at RIC) investigates the sensory-motor system through a close interaction with artificial systems. Specifically, we are interested in determining how the brain acquires, organizes and executes motor behaviors. We use robotic and interface technologies to investigate how humans adapt to radical changes in the environment and in body mechanics.
Consistent evidence indicates that the nervous system is capable of coping with changes in the body and in the environment by developing internal representations of the relationship between movement commands and their sensory consequences. In this sense, motor learning is not only about improving performance. Motor learning is a means by which our brain develops an understanding of the physical and statistical properties of the world. We are studying the basic properties of this learning process and how it may be exploited to facilitate rehabilitation. Other studies within our group are directed at facilitating bidirectional communications between the human body and artificial instruments, such as wheelchairs and computers. We wish to combine the biological mechanisms of learning with machine learning algorithms for reducing the burden that disabled people must currently endure for the efficient operation of systems such as powered wheelchairs and other assistive devices. In a nutshell: we want to create systems that learn and adapt to their users.
Understanding how the brain controls motor behavior is of clinical interest since alterations in neuromotor control due to stroke and other neurological impairments can severely limit motor function. Through our research we wish to create knowledge that can help restore motor functions in individuals with neurological disorders.
For lab information and more, see Dr. Mussa-Ivaldi's faculty profile.
See Dr. Mussa-Ivaldi's publications on PubMed.
Contact Dr. Mussa-Ivaldi at 312-238-1230 or the Robotics Lab at 312-238-1232.
Understanding the computational implications of neural dynamics
The goal of our research is to understand information processing in the brain. We use mathematical models based on specific hypothesis about encoding and decoding aspects of neural activity, and use analytical and numerical techniques to investigate the implications of these hypothesis so that they can be validated, modified, or discarded as dictated by experimental data.
Our purpose is to understand the computational implications of neural dynamics. Our work relies on conceptual frameworks and mathematical tools from statistical physics, information theory, nonlinear dynamics, probability theory, and machine learning, and aims at formulating data driven models that illuminate specific aspects of information processing by networks of neurons.
Specific topics of interest include input-output characteristics of single cell and network models, encoding and decoding of information through neural activity, early stages of sensory processing, and the neural control of movement. We work in close collaboration with experimental groups, both at Northwestern University and at other institutions. Recently, we have focused on the interplay between neural connectivity, network dynamics, and computation, and on brain-machine interfaces for the decoding of neural activity in motor cortex and the encoding of sensory information via stimulation of somatosensory cortex. Our work on brain-machine interfaces is funded by NINDS, the National Institute of Neurological Disorders and Strokes within the NIH.
For lab information and more, see Dr. Solla’s faculty profile.
See Dr. Solla's publications on PubMed.
Contact Dr. Solla at 312-503-1408 or the lab at 312-503-1408.
Understanding the principles of neuronal dysfunction in disease states
Our group has five research topics. The first topic area is what drives Parkinson’s disease (PD). Using a combination of optical, electrophysiological and molecular approaches, we are examining the factors governing neurodegeneration in PD and its network consequences, primarily in the striatum. This work has led to a Phase III neuroprotection clinical trial for early stage PD and a drug development program targeting a sub-class of calcium channels. The second topic area is network dysfunction in Huntington’s disease (HD). Using the same set of approaches, we are exploring striatal and pallidal dysfunction in genetic models of HD, again with the aim of identifying novel drug targets. The third topic area is striatal dysfunction in schizophrenia, with a particular interest in striatal adaptations to neuroleptic treatment. The fourth topic area is post-traumatic stress disorder and the role played by neurons in the locus ceruleus in its manifestations. The last topic area is chronic pain states and the impact these have on the circuitry of the ventral striatum.
For lab information and more, see Dr. Surmeier's faculty profile.
See Dr. Surmeier's publications on PubMed.
Contact Dr. Surmeier at 312-503-4904.
Yijuan Du, Patricia Gonzalez Rodriguez, Steven Graves, Martin Henrich, Ema Ilijic, Harini Lakshminarasimhan, Austin Lim, Stephen Logan, Curtis Neveu, Tamara Perez-Rosello, DeNard Simmons, Asami Tanimura, Tatiana Tkatch, Cecilia Tubert, Sasha Ulrich, Enrico Zampese, Shenyu Zhai