Cosyne 2009 Workshops
March 2-3, 2009
Snow Bird, Utah
Workshop Title
Modulation of cortical responses by behavior and brain state
Organizer(s)
Robert C. Froemke, UCSF (primary contact)
Michael DeWeese, University of California, Berkeley
Abstract
Information processing by cortical networks is influenced by many factors beyond the functional organization of local circuitry, including behavioral context, arousal level, and neuromodulatory cues from subcortical nuclei. The overall goal of this workshop is to understand how cortical dynamics are modulated by these extrinsic signals and governed by behavioral states. Recent work using a range of experimental techniques, including intracellular recordings in vivo and recordings from behaving animals, have produced a wealth of new data under different conditions; we aim in this workshop to synthesize these results into a coherent picture of cortical modulation. There are three basic questions to be addressed:
- How behavioral task demands affect cortical responses to sensory inputs
- How cortical activity changes under different global states (sleep, arousal, anesthesia)
- Which brain areas and neuromodulators are involved in control of cortical dynamics
Schedule
Morning session (8:00 – 11:00AM)
8:00 – 8:20 Jonathon Fritz Multiple forms of attentional modulation at different cortical levels
8:30 – 8:50 Arthur Houweling Nanostimulation: manipulation of single neuron activity in the juxtacellular configuration
9:00 – 9:20 Vlad Vyazovskiy Cellular correlates of sleep homeostasis in the rat
9:30 – 10:00 Break for coffee and discussion
10:00 - 10:20 Pascal Fries A microsaccadic rhythm modulates gamma-band synchronization and behavior
10:30 – 10:50 Santiago Jaramillo Temporal expectation modulates neuronal responses in auditory cortex
Afternoon session (4:30 – 7:30PM)
4:30 – 4:50 Shih-Chieh Lin Attentional modulation of cortical ERPs gated by basal forebrain ensemble bursting
5:00 – 5:20 Stephen David Effect of behavior state on cortical representation and routing of auditory information
5:30 – 5:50 Martin Sarter Getting the attention of cortical networks
6:00 - 6:30 Break for coffee & discussion
6:30 – 6:50 Alex Thiele Mechanisms of attention in V1
7:00 – 7:20 Jude Mitchell Attention reduces low frequency correlated noise and burst firing in macaque V4
Speakers
1. Jonathon Fritz, Stephen David, Pingbo Yin, Daniel Winkowski & Shihab Shamma, University of Maryland, Multiple forms of attentional modulation at different cortical levels: adaptive receptive field plasticity in primary sensory cortex, and top-down modulatory signals and behavioral gating in prefrontal cortex
Our previous behavioral physiology studies of attentionally-driven, task-related plasticity in primary auditory cortex (A1) have shown the spectro-temporal receptive fields (STRFs) of A1 neurons rapidly adapt to current task-relevant spectral and temporal cues in auditory tasks. By taking multiple on-line STRF “snapshots” of A1 neurons during behavior (using reverse correlation techniques), we can measure dynamic STRF shape changes in real time while ferrets perform either spectral, temporal or a series of both types of tasks. In order to explore the broader neural network involved in task-related plasticity, we have begun recording from other cortical areas, including auditory association cortex and prefrontal cortex. Since one of the defining features of neurons in the prefrontal cortex (PFC) is selective encoding of task-relevant information we recorded dorsal PFC responses in non-task, quiescent states and during multiple auditory and visual task conditions in behaving head-fixed ferrets. We observed responses that were highly adaptive, selective, and categorical and often gated by behavior. Our results suggest that the PFC rapidly resets responses in a state-dependent manner, tracking attended task-relevant information in changing task conditions. However, in passive non-task conditions, responses to incoming information may be shaped by a fading memory of the behaviorally relevant target from the most recent task. Since the time course of acquisition and extinction of target representations in PFC paralleled the temporal trajectory for task-related receptive field plasticity in A1, primary auditory cortex (Fritz et al., 2003, 2007), we have also begun simultaneous recording in PFC and A1. Our results suggest that PFC contributes to top-down modulation of A1 responses. Moreover, we have recently found that PFC stimulation in conjunction with paired pure tones, leads to receptive field plasticity in A1, similar to the plasticity observed in associative pairing of tones with stimulation of nucleus basalis. We conjecture that PFC may orchestrate changes in sensory cortex during behavior by direct top-down projections and also indirectly, by influencing the activity of key subcortical neuromodulators. Thus, attention may act differentially to modulate multiple cortical and subcortical levels to enhance behavior.
2. Arthur Houweling & Michael Brecht, Humboldt University, Nanostimulation: manipulation of single neuron activity in the juxtacellular configuration
In the mammalian brain thousands of single neuron recording studies have been performed but only a handful of single-cell stimulation studies. This paucity of single-cell stimulation data reflects a lack of easily applicable single-cell stimulation techniques. We provide an overview of the procedures involved in nanostimulation, a methodology derived from the juxtacellular labeling technique. Nanostimulation is easy to apply and can be directed to a wide variety of identifiable neurons in anesthetized and awake animals. To obtain the juxtacellular configuration we use glass pipettes with a DC resistance of ~ 5 MOhm and closely approach a neuron until we observe a several-fold increase in resistance to values > 20 MOhm. The recorded AP amplitude grows to > 2 mV and neurons can be activated with currents in the nanoampere range – hence the term nanostimulation. While exact AP timing has not been achieved, AP frequency and AP number can be parametrically controlled. We demonstrate that nanostimulation can also be used to selectively inhibit sensory responses in identified neurons. There is strong evidence to suggest that nanostimulation selectively activates single neurons and that effects are cell-specific. Nanostimulation holds therefore great potential for elucidating how single neurons contribute to behavior.
3. Vlad Vyazovskiy (1), Umberto Olcese (1,2), Yaniv M. Lazimy (1), Ugo Faraguna (1), Steve K. Esser (1), Justin C. Williams (2), Chiara Cirelli (1) & Giulio Tononi (1), 1) Department of Psychiatry, University of Wisconsin-Madison, 2) Scuola Superiore Sant'Anna, Pisa, Italy, Cellular correlates of sleep homeostasis in the rat
Slow-wave activity (SWA) is defined as the EEG power in the 0.5-4.0 Hz range during NREM sleep. SWA ultimately reflects the number and amplitude of slow waves during NREM sleep, and is strongly influenced by the duration of preceding wakefulness. Specifically, SWA is high in early sleep (high sleep pressure) and decrease across the sleep period, when sleep pressure is released, and is higher at sleep onset after sleep deprivation. However, the cellular mechanisms underlying these SWA changes are unclear. We hypothesized that the homeostatic evolution of SWA reflects specific changes in neuronal firing patterns. We employed combined recordings of the EEG and cortical multiunit activity and investigated their relationship across the sleep-wake cycle in freely behaving rats. High-amplitude negative EEG slow waves during NREM sleep coincided with the occurrence of population silent periods preceded and followed by periods of elevated neuronal activity. Such periods reflect the slow oscillation of individual neurons that consists of a depolarized up state, characterized by sustained irregular firing, and a hyperpolarized down state, characterized by neuronal silence. We found that: i) the firing of most cortical neurons under high sleep pressure is modulated by the phase of high-amplitude EEG slow waves; ii) dissipation of sleep pressure results in progressive uncoupling of individual neurons from the population activity, and in loss of their phase-relationship with the EEG slow waves; iii) in the course of sleep an increasing fraction of neurons fire out of phase with the rest of the population, resulting in longer periods of sustained population activity due to shortening of the population silent periods; iv) the decreased incidence of long silent periods results in reduced occurrence of high-amplitude slow waves and progressive homeostatic decline of SWA. Finally, compared to low sleep pressure, high sleep pressure is associated with elevated neuronal firing rates during the active periods of NREM sleep, as well as during REM sleep and waking. In summary, cortical neuronal activity patterns are affected by the sleep-wake history. Specifically, these results suggest that the global homeostatic decline in SWA across sleep arises from progressive reduction of synchronous firing among individual neurons.
4. Conrado A. Bosman (1), Thilo Womelsdorf (1), Robert Desimone (2,3), Pascal Fries (1), 1) Donders Institute for Brain, Cognition and Behaviour, Radboud University, The Netherlands, 2) Laboratory of Neuropsychology, National Institute of Mental Health, 3) McGovern Institute for Brain Research at the Massachusetts Institute of Technology, A microsaccadic rhythm modulates gamma-band synchronization and behavior
Rhythms occur both in neuronal activity and in behavior. Behavioral rhythms abound at frequencies at or below 10 Hz. Neuronal rhythms cover a very wide frequency range, and the phase of neuronal low frequency rhythms often rhythmically modulates the strength of higher frequency rhythms, particularly of gamma-band synchronization (GBS). We studied stimulus induced GBS in awake monkey areas V1 and V4 in relation to a specific form of spontaneous behavior, namely microsaccades (MSs), small fixational eye movements. We found that MSs occur rhythmically at a frequency of about 3.3 Hz. The rhythmic MSs were predicted by the phase of the 3.3 rhythm in V1 and V4 local field potentials (LFP). In turn, the MSs modulated both, visually induced GBS and the speed of visually triggered behavioral responses. Fast/slow responses were preceded by a specific temporal pattern of MSs. These MS patterns induced perturbations in GBS that in turn explained variability in behavioral response speed. We hypothesize that the 3.3 Hz rhythm structures the sampling and exploration of the environment through building and breaking neuronal ensembles synchronized in the gamma-frequency band to process sensory stimuli.
5. Santiago Jaramillo & Anthony M. Zador, Cold Spring Harbor Laboratory, Temporal expectation modulates neuronal responses in auditory cortex
When a stimulus occurs at a predictable instant in time, anticipation of the stimulus improves the speed and accuracy with which it is detected. We have developed a two-alternative choice task paradigm in freely moving rats to study the neural mechanisms underlying this phenomenon in the auditory system. Behavioral measurements confirmed that valid expectations improved both reaction times and detection thresholds. We then used tetrodes to record responses from single neurons in the auditory cortex of rats performing the task. Responsive neurons often showed an increase evoked response to tones immediately preceding the expected moment of appearance of the target when compared against responses to the same tones occurring long before the expected target. In addition, grouping behavioral trials according to the subject's reaction time revealed correlations between the strength of the neuronal responses and performance.
6. Shih-Chieh Lin & Miguel A. L. Nicolelis, Duke University Medical Center, Attentional modulation of cortical ERPs gated by basal forebrain ensemble bursting
One important electrophysiological index of attention is the enhancement of cortical event-related potentials (ERPs). Yet it remains unclear what additional neural systems are recruited to mediate the influences of attention on ERPs. Here we investigate a novel hypothesis that attention-related cortical ERPs may be gated by a subcortical mechanism: ensemble bursting of non-cholinergic basal forebrain (BF) neurons. We show that, in rats performing an oddball task, one component of prefrontal cortex (PFC) ERP was tightly correlated with the strength and timing of BF ensemble bursting in single trials. Furthermore, the PFC ERP response displayed a characteristic layer-specific depth profile independent of the stimulus sensory modality. These results support the conclusion that attention-related cortical ERPs are likely gated by ensemble bursting of non-cholinergic BF neurons, which can convert the motivational salience of attended stimuli (Lin and Nicolelis, Neuron 2008) into transient enhancement of cortical activity (Lin et al, J Neurophysiol 2006).
7. Stephen V. David, Nima Mesgarani, Jonathan B. Fritz & Shihab A. Shamma, University of Maryland, Effect of behavior state on cortical representation and routing of auditory information
Auditory behavior requires classifying sounds according to spectro-temporal properties and their associated meaning. In simple operant conditioning paradigms, one class of sounds (targets) requires an active response while another class (distracters) requires no change in behavior. The prevailing neurophysiological model for such behaviors is that sensory cortex operates as a matched filter, enhancing the responses of neurons that signal targets relative to distracters. However, recent studies have shown that additional contingencies, including reward value, motivation and task difficulty, can also modulate sensory responses. To learn more about the influence of task contingencies during sound discrimination, we recorded from neurons in ferret primary auditory cortex (A1) and prefrontal cortex (PFC) under two behavior paradigms. Both tasks required the same auditory discrimination, detecting a pure tone target in a sequence of broadband noise bursts. The first paradigm used conditioned avoidance: Water reward was delivered during distracters, and subjects were punished with a small shock if they failed to pause drinking after a target. The second paradigm used positive reinforcement: Target hits were rewarded with water, and false alarms were punished with a timeout.
We compared the spectro-temporal tuning properties of single A1 neurons during both tasks and also during passive listening. During conditioned avoidance, neurons showed enhanced responses to the target tone frequency. During positive reinforcement, on the other hand, A1 neurons showed an opposite pattern of modulation, with a decreased relative response to the target frequency. Responses of single neurons in PFC were strongly gated by behavior and also varied between behavior paradigms. During conditioned avoidance, neurons responded only to target sounds, while during positive reinforcement they also responded to reference sounds. In both cases, PFC responses were strong for the stimuli whose representation was relatively enhanced in A1. These results are consistent with a generalized matched filter model in which PFC responses feed back to A1 and enhance responses to stimuli that signal the suppression of basal appetitive behaviors. Recordings of local field potentials suggest a gating mechanism by which task-relevant information is propagated from A1 to PFC.
8. Martin Sarter, University of Michigan, Getting the attention of cortical networks
Conventional descriptions of the role of neuromodulator systems in attention have suggested that neuromodulators act relatively slowly (on the scale of tens of seconds of even minutes) to influence ³arousal² states and gate input processing in the cortex. Results from conventional methods available to monitor the activity of neuromodulator systems performing animals inadvertently substantiated such views. Our recent research using a new electrochemical technique to measure the release of acetylcholine at a sub-second resolution in performing animals indicated that second-based cholinergic transients in the prefrontal cortex mediate, and are necessary for, fundamental cognitive operations such as cue detection and processing mode shifts. These results reject the conventional description of the cortical cholinergic input system as a neuromodulator system and instead suggest that cholinergic transients determine the state of cortical circuits on a scale of seconds and thereby control attentional performance.
9. Alex Thiele, Newcastle University, Mechanisms of attention in V1
Attention exerts a strong influence over neuronal processing in cortical areas. It selectively increases firing rates and affects tuning properties, including changing receptive field locations and sizes. While these effects are well studied, their cellular mechanisms are poorly understood. To study these, we combined iontophoretic pharmacological analysis of cholinergic receptors with single cell recordings in V1 while macaque monkeys performed a task that demanded top-down attention. Attending to the receptive field of the V1 neuron under study caused an increase in firing rates. This attentional modulation was enhanced by low doses of acetylcholine. Furthermore, applying the muscarinic antagonist scopolamine reduced attentional modulation, while the nicotinic antagonist mecamylamine had no systematic effect. These results demonstrate that muscarinic cholinergic mechanisms play a central role in mediating the effects of attention in V1, while nicotinic receptors in V1 seem mostly involved in modulation of sensory transmission.
10. Jude Mitchell & John H. Reynolds, The Salk Institute, La Jolla, Attention reduces low frequency correlated noise and burst firing in macaque V4
Although previous studies of attention often focus on increases in firing rate as the neural correlate of improved processing for attended items, it is also important to consider how attention alters the variability of response. In particular, if sources of variability are correlated at the population level then it is not possible to average out noise by pooling, thus limiting the accuracy with which sensory information can be decoded (Zohary et. al., 1994). Recently we found that attending to a sustained stimulus inside a V4 neuron's receptive field led to a reduction in the trial-to-trial variability of the response, as measured by the Fano Factor, and that reductions were much stronger among neurons with narrower action potentials, putative fast spiking interneurons (Mitchell et. al., 2007). Here we examine the timescales of the underlying firing rate fluctuations that contribute to this variability, how they differ between neuronal classes, and whether or not they are correlated at the population level. The dominant source of noise in both neuronal classes is due to low frequency (< 8hz) fluctuations in firing rate. Broad spiking neurons, putative pyramids, also exhibit a second source of noise at short timescales that is due to firing in bursts, doublets or triplets of spikes, and is also reduced by attention. We analyzed the correlations in spiking of simultaneously recorded pairs to evaluate if these sources of noise were correlated in the population. We find that the low frequency rate fluctuations are correlated across units and importantly that these correlations are reduced by attention. Low frequency fluctuations are ubiquitous to cortical firing and may represent one source of recurrent network noise that attention modulates to improve sensory signals.
