Final Report Summary - WHISKERATTENTION (The effects of spatial attention on stimulus encoding in rat barrel cortex and on relationships between barrel cortex and locus coeruleus norepinephrine neuron spiking)
Summary of the project objectives
Research Objective 1: Development of a rodent attention task
The aim of this objective was to develop a model of top-down attention in rats. Top-down attention describes an organism’s use of associations between sensory stimuli and rewards to selectively enhance the representation of reward-predicting stimuli in the brain. Although cognitive neuroscience studies with rats commonly use operant chambers or mazes in which the rat is free to move, we created a paradigm for the head-fixed rat. Head-fixation provided a high degree of control over sensory stimulation and behavioral monitoring, which was predicted to decrease variability and to allow more detailed analysis of behavior and brain activity. Head-fixation also provides an opportunity to video-track pupil diameter, which correlates with arousal and attention. Our ultimate goal was to study the effects of attention to left and right space by training the rat to associate reward with deflections of whiskers while a preceding auditory cue would either validly or invalidly instruct them to expect the whisker stimulus on one side. Developing a complex and novel behavioral task, and employing novel methods (such as head-fixation in rats), present technical issues and, to mitigate this risk, we proposed a back-up behavioral task. The back-up task used a simple, well-explored behavioral, called extinction, in which the association between reward and stimulus is extinguished and becomes unattended.
Research Objectives 2 and 3: Measure the effects of attention on stimulus encoding in specific cortical layers and on communication between cortical layers
Physiological studies of attention have demonstrated that sensory neurons represent attended stimuli so that the neural representation is more informative, stronger magnitude, and more efficiently communicated to brain regions controlling decisions and actions concerning the stimulus. Anatomical studies have demonstrated that attentional modulation of sensory neurons should originate from the prefrontal cortex (PFC) and that the modulation should occur in deep (layer 6) of sensory cortex. Our goal was to record activity from individual cortical layers. These data would be analyzed for layer-specific effects of attention on the information content, magnitude, and communication efficiency of cortical neuron activity evoked by the attended stimulus.
Micro-electrode recordings from the head-fixed rat required two methodological advances. Data collection first required developing a stable, implanted recording chamber that could accommodate the head-fixation post and allowed targeting of a wide array of brain regions. Secondly, due to the position of some brain regions in relation to skull and chamber, novel electrode designs and histological methods were required for recording laminar activity and verifying electrode tip location, respectively.
Research Objective 4: Measure the effects of attention on the relationship between the norepinephrine-producing locus coeruleus and the sensory cortex
Medications which effect norepinephrine (NE) can improve attention in humans and animals and NE is released in both sensory cortex and in the PFC, which provides signals to sensory cortex that modulate the representation of attended stimuli. Our goal was to record neurons in the locus coeruleus (LC), which produces NE and to correlate LC activity with cortical activity during attention.
Description of the work performed since the beginning of the project
Research Objective 1
We developed a stable head-fixation post and recording chamber. Using this preparation, we initially trained head-fixed rats to run on a spherical treadmill in order to obtain reward. We tested two methods of whisker deflection as sensory stimuli, but each presented technical challenges, which we could not resolve. However, subsequent testing with visual and auditory stimuli revealed that rats would respond to both types of stimulation. Therefore, we succeeded in developing a complex attention task for the rat which required attending to either the auditory or visual component of a compound auditory-visual stimulus.
Research Objectives 2 and 3
We have recorded brain activity during the task using microelectrodes implanted in sensory cortex (auditory cortex and visual cortex) and PFC as attention is shifted between the auditory and visual aspects of the stimulus. The analysis is ongoing, but initial findings of differences in beta (12 – 20 Hz) frequency power and coherence of the local field potential (LFP) across regions suggests that attention affects communication between PFC and sensory cortex depending on which sensory modality is attended. Due to the medial position of the PFC in rats and mice, it is uncommon to simultaneously record activity from multiple PFC layers. Therefore, we developed a PFC laminar electrode array and histological methods for verifying electrode tip location.
Research Objective 4
Lesion studies have previously demonstrated that LC and PFC are both required for attention. We began studying interactions between the LC and PFC by recording both structures simultaneously, including sampling all PFC layers, under anesthesia. Electrical stimulation was used in each region to examine the effects of each structure on activity in the downstream structure. Our analysis revealed strong top-down modulation from PFC to the LC. We are currently testing whether the PFC-controlled LC neurons project to sensory cortex, motor cortex, or thalamus, which may have implications for PFC control over NE release in sensory cortex during attention to behaviorally-relevant stimuli. We also observed that LC has complex effects of layer specific population activity, spike timing, and communication between layers in PFC. Our next objective is to record these structures during the behavioral task. Finally, by using a similar rat attention task in an operant chamber, we demonstrated that increasing brain-wide NE accelerated attending to a new reward-predicting stimulus. The compound used to increase NE is a clinical compound used to ameliorate cognitive impairment in humans (Atomoxetine), thus demonstrating the usefulness of rodent models of human attention.
Description of the main results achieved so far
• Attention modulates stimulus-evoked beta oscillations in the PFC and sensory cortex. Ongoing work: After collecting additional data, we will analyze layer-specific coherence and Granger causality between LFP in PFC and sensory cortex. Attentional modulation of stimulus-evoked single unit spiking in PFC, sensory cortex, and LC will be examined for phase locking to the multi-region coherent LFP, which may enhance communication between these regions. Pupil diameter, which correlates with arousal and LC activity, will be used to measure attention-related NE changes.
• Increased systemic NE accelerates shifting attention (Totah, Logothetis, and Eschenko. 2015. Submitted).
• LC stimulation effects LFP-LFP and LFP-spike gamma (40 – 60 Hz and 60 – 100 Hz) synchrony heterogeneously depending on cortical layer. The effect occurs after neuronal excitation in thalamo-cortical recipient layers. Ongoing work: Inactivation of the thalamus using micro-injections of the GABA-a agonist muscimol will demonstrate if LC modulation occurs via thalamic terminals in PFC. Granger causality between PFC layers will provide novel information about laminar network activity in the PFC, which has rarely been studied with laminar electrodes.
• PFC stimulation excites specific LC single neurons, which may form specialized circuits for top-down (PFC) control over sensory cortex NE. Ongoing work: the connections of individual LC neurons are being traced to multiple cortical and sub-cortical sites.
Expected final results and their potential impact and use
We have submitted one paper (Totah, Logothetis, Eschenko. 2015. Atomoxetine accelerates attentional set shifting without effecting learning rate in the rat). After ongoing work is completed, we will submit two additional papers to high-impact neuroscience journals which cover the effects of LC and PFC interactions with one study on LC stimulation effects on PFC neural activity and the other on PFC stimulation effects on LC neural activity. Using the support of the Marie Curie Actions, we were able to fully develop a novel attention paradigm and the capability for multi-site laminar recordings of cortex and LC. The analytical methods developed for studying LC-cortex interactions in anesthetized experiments will be applied to recordings during the attention task after data collection is completed. We expect recordings of neural activity and pupil diameter during attention to provide novel insight into NE modulation of cortical interactions and the relation between the pupil, attention, and NE. Finally, the novel rodent model will provide a springboard our future studies using optogenetics in mice and rats to conduct targeted modulation of the NE system.
Research Objective 1: Development of a rodent attention task
The aim of this objective was to develop a model of top-down attention in rats. Top-down attention describes an organism’s use of associations between sensory stimuli and rewards to selectively enhance the representation of reward-predicting stimuli in the brain. Although cognitive neuroscience studies with rats commonly use operant chambers or mazes in which the rat is free to move, we created a paradigm for the head-fixed rat. Head-fixation provided a high degree of control over sensory stimulation and behavioral monitoring, which was predicted to decrease variability and to allow more detailed analysis of behavior and brain activity. Head-fixation also provides an opportunity to video-track pupil diameter, which correlates with arousal and attention. Our ultimate goal was to study the effects of attention to left and right space by training the rat to associate reward with deflections of whiskers while a preceding auditory cue would either validly or invalidly instruct them to expect the whisker stimulus on one side. Developing a complex and novel behavioral task, and employing novel methods (such as head-fixation in rats), present technical issues and, to mitigate this risk, we proposed a back-up behavioral task. The back-up task used a simple, well-explored behavioral, called extinction, in which the association between reward and stimulus is extinguished and becomes unattended.
Research Objectives 2 and 3: Measure the effects of attention on stimulus encoding in specific cortical layers and on communication between cortical layers
Physiological studies of attention have demonstrated that sensory neurons represent attended stimuli so that the neural representation is more informative, stronger magnitude, and more efficiently communicated to brain regions controlling decisions and actions concerning the stimulus. Anatomical studies have demonstrated that attentional modulation of sensory neurons should originate from the prefrontal cortex (PFC) and that the modulation should occur in deep (layer 6) of sensory cortex. Our goal was to record activity from individual cortical layers. These data would be analyzed for layer-specific effects of attention on the information content, magnitude, and communication efficiency of cortical neuron activity evoked by the attended stimulus.
Micro-electrode recordings from the head-fixed rat required two methodological advances. Data collection first required developing a stable, implanted recording chamber that could accommodate the head-fixation post and allowed targeting of a wide array of brain regions. Secondly, due to the position of some brain regions in relation to skull and chamber, novel electrode designs and histological methods were required for recording laminar activity and verifying electrode tip location, respectively.
Research Objective 4: Measure the effects of attention on the relationship between the norepinephrine-producing locus coeruleus and the sensory cortex
Medications which effect norepinephrine (NE) can improve attention in humans and animals and NE is released in both sensory cortex and in the PFC, which provides signals to sensory cortex that modulate the representation of attended stimuli. Our goal was to record neurons in the locus coeruleus (LC), which produces NE and to correlate LC activity with cortical activity during attention.
Description of the work performed since the beginning of the project
Research Objective 1
We developed a stable head-fixation post and recording chamber. Using this preparation, we initially trained head-fixed rats to run on a spherical treadmill in order to obtain reward. We tested two methods of whisker deflection as sensory stimuli, but each presented technical challenges, which we could not resolve. However, subsequent testing with visual and auditory stimuli revealed that rats would respond to both types of stimulation. Therefore, we succeeded in developing a complex attention task for the rat which required attending to either the auditory or visual component of a compound auditory-visual stimulus.
Research Objectives 2 and 3
We have recorded brain activity during the task using microelectrodes implanted in sensory cortex (auditory cortex and visual cortex) and PFC as attention is shifted between the auditory and visual aspects of the stimulus. The analysis is ongoing, but initial findings of differences in beta (12 – 20 Hz) frequency power and coherence of the local field potential (LFP) across regions suggests that attention affects communication between PFC and sensory cortex depending on which sensory modality is attended. Due to the medial position of the PFC in rats and mice, it is uncommon to simultaneously record activity from multiple PFC layers. Therefore, we developed a PFC laminar electrode array and histological methods for verifying electrode tip location.
Research Objective 4
Lesion studies have previously demonstrated that LC and PFC are both required for attention. We began studying interactions between the LC and PFC by recording both structures simultaneously, including sampling all PFC layers, under anesthesia. Electrical stimulation was used in each region to examine the effects of each structure on activity in the downstream structure. Our analysis revealed strong top-down modulation from PFC to the LC. We are currently testing whether the PFC-controlled LC neurons project to sensory cortex, motor cortex, or thalamus, which may have implications for PFC control over NE release in sensory cortex during attention to behaviorally-relevant stimuli. We also observed that LC has complex effects of layer specific population activity, spike timing, and communication between layers in PFC. Our next objective is to record these structures during the behavioral task. Finally, by using a similar rat attention task in an operant chamber, we demonstrated that increasing brain-wide NE accelerated attending to a new reward-predicting stimulus. The compound used to increase NE is a clinical compound used to ameliorate cognitive impairment in humans (Atomoxetine), thus demonstrating the usefulness of rodent models of human attention.
Description of the main results achieved so far
• Attention modulates stimulus-evoked beta oscillations in the PFC and sensory cortex. Ongoing work: After collecting additional data, we will analyze layer-specific coherence and Granger causality between LFP in PFC and sensory cortex. Attentional modulation of stimulus-evoked single unit spiking in PFC, sensory cortex, and LC will be examined for phase locking to the multi-region coherent LFP, which may enhance communication between these regions. Pupil diameter, which correlates with arousal and LC activity, will be used to measure attention-related NE changes.
• Increased systemic NE accelerates shifting attention (Totah, Logothetis, and Eschenko. 2015. Submitted).
• LC stimulation effects LFP-LFP and LFP-spike gamma (40 – 60 Hz and 60 – 100 Hz) synchrony heterogeneously depending on cortical layer. The effect occurs after neuronal excitation in thalamo-cortical recipient layers. Ongoing work: Inactivation of the thalamus using micro-injections of the GABA-a agonist muscimol will demonstrate if LC modulation occurs via thalamic terminals in PFC. Granger causality between PFC layers will provide novel information about laminar network activity in the PFC, which has rarely been studied with laminar electrodes.
• PFC stimulation excites specific LC single neurons, which may form specialized circuits for top-down (PFC) control over sensory cortex NE. Ongoing work: the connections of individual LC neurons are being traced to multiple cortical and sub-cortical sites.
Expected final results and their potential impact and use
We have submitted one paper (Totah, Logothetis, Eschenko. 2015. Atomoxetine accelerates attentional set shifting without effecting learning rate in the rat). After ongoing work is completed, we will submit two additional papers to high-impact neuroscience journals which cover the effects of LC and PFC interactions with one study on LC stimulation effects on PFC neural activity and the other on PFC stimulation effects on LC neural activity. Using the support of the Marie Curie Actions, we were able to fully develop a novel attention paradigm and the capability for multi-site laminar recordings of cortex and LC. The analytical methods developed for studying LC-cortex interactions in anesthetized experiments will be applied to recordings during the attention task after data collection is completed. We expect recordings of neural activity and pupil diameter during attention to provide novel insight into NE modulation of cortical interactions and the relation between the pupil, attention, and NE. Finally, the novel rodent model will provide a springboard our future studies using optogenetics in mice and rats to conduct targeted modulation of the NE system.