Для просмотра этого контента требуется подписка на Jove Войдите в систему или начните бесплатную пробную версию.

Резюме

This experiment uses an anatomically-constrained magnetoencephalography (aMEG) method to examine brain oscillatory dynamics and long-range functional synchrony during engagement of cognitive control as a function of acute alcohol intoxication.

Abstract

Decision making relies on dynamic interactions of distributed, primarily frontal brain regions. Extensive evidence from functional magnetic resonance imaging (fMRI) studies indicates that the anterior cingulate (ACC) and the lateral prefrontal cortices (latPFC) are essential nodes subserving cognitive control. However, because of its limited temporal resolution, fMRI cannot accurately reflect the timing and nature of their presumed interplay. The present study combines distributed source modeling of the temporally precise magnetoencephalography (MEG) signal with structural MRI in the form of "brain movies" to: (1) estimate the cortical areas involved in cognitive control ("where"), (2) characterize their temporal sequence ("when"), and (3) quantify the oscillatory dynamics of their neural interactions in real time. Stroop interference was associated with greater event-related theta (4 – 7 Hz) power in the ACC during conflict detection followed by sustained sensitivity to cognitive demands in the ACC and latPFC during integration and response preparation. A phase-locking analysis revealed co-oscillatory interactions between these areas indicating their increased neural synchrony in theta band during conflict-inducing incongruous trials. These results confirm that theta oscillations are fundamental to long-range synchronization needed for integrating top-down influences during cognitive control. MEG reflects neural activity directly, which makes it suitable for pharmacological manipulations in contrast to fMRI that is sensitive to vasoactive confounds. In the present study, healthy social drinkers were given a moderate alcohol dose and placebo in a within-subject design. Acute intoxication attenuated theta power to Stroop conflict and dysregulated co-oscillations between the ACC and latPFC, confirming that alcohol is detrimental to neural synchrony subserving cognitive control. It interferes with goal-directed behavior that may result in deficient self-control, contributing to compulsive drinking. In sum, this method can provide insight into real-time interactions during cognitive processing and can characterize the selective sensitivity to pharmacological challenge across relevant neural networks.

Introduction

The overall goal of this study is to examine the effects of acute alcohol intoxication on spatio-temporal changes in the brain oscillatory dynamics and long-range functional integration during cognitive control. The employed multimodal imaging approach combines magnetoencephalography (MEG) and structural magnetic resonance imaging (MRI) to provide insight into the neural basis of decision making with high temporal precision and at the level of an interactive system.

Flexible behavior makes it possible to adapt to changing contextual demands and to switch strategically between different tasks and requirements in agreement with one's intents and goals. The capacity to suppress automatic responses in favor of goal-relevant but non-habitual actions is an essential aspect of cognitive control. Extensive evidence suggests that it is subserved by a predominantly frontal cortical network, with the anterior cingulate cortex (ACC) as a central node in this interactive network^1,2,3,4. While the abundant anatomical connectivity between the ACC and lateral frontal cortices is well-described^5,6, the functional characteristics of communication between these regions during cognitive control, response selection and execution, are poorly understood.

The highly influential conflict monitoring theory^7,8 proposes that cognitive control arises from a dynamic interaction between the medial and lateral prefrontal cortices. This account purports that the ACC monitors conflict between competing representations and engages the lateral prefrontal cortex (latPFC) to implement response control and optimize performance. However, this account is primarily based on the functional MRI (fMRI) studies using the blood oxygenation level dependent (BOLD) signal. The fMRI-BOLD signal is an excellent spatial mapping tool, but its temporal resolution is limited because it reflects regional hemodynamic changes mediated by neurovascular coupling. As a result, the BOLD signal changes unfold on a much slower time scale (in seconds) than the underlying neural events (in milliseconds)⁹. Moreover, the BOLD signal is sensitive to alcohol's vasoactive effects¹⁰ and may not accurately represent the magnitude of neural changes, which makes it less suitable for studies of acute alcohol intoxication. Therefore, the presumed interplay between the medial and lateral prefrontal cortices and its sensitivity to alcohol intoxication need to be examined by methods that record neural events in a temporally precise manner. MEG has an excellent temporal resolution since it directly reflects postsynaptic currents. The anatomically-constrained MEG (aMEG) methodology employed here is a multimodal approach that combines distributed source modeling of the MEG signal with structural MRI. It allows for the estimation of where the conflict- and beverage-related brain oscillatory changes are occurring and to understand the temporal sequence ("when") of the involved neural components.

Decision making relies on the interactions of distributed brain regions that are dynamically engaged to deal with increased demands on cognitive control. One way to estimate event-related changes in long-range synchrony between two cortical regions is to calculate their phase coupling as an index of their co-oscillations^11,12. The present study applied a phase-locking analysis to test the basic tenet of the conflict monitoring theory by examining the co-oscillatory interactions between the ACC and latPFC. Neural oscillations in theta range (4 – 7 Hz) are associated with cognitive control and have been proposed as a fundamental mechanism supporting the long-range synchronization needed for top-down cognitive processing^13,14,15,16. They are generated in prefrontal areas as a function of task difficulty and are significantly attenuated by acute alcohol intoxication^17,18,19,20.

Long-term excessive alcohol intake is associated with a range of cognitive deficits with prefrontal circuitry being especially affected^21,22. Acute alcohol intoxication is detrimental to cognitive control under conditions of increased difficulty, ambiguity, or those that induce response incompatibility^17,23,24. By affecting decision making, alcohol may interfere with goal-directed behavior, may result in poor self-control and increased drinking, and may also contribute to traffic- or work-related hazards^25,26,27. The present study uses an aMEG approach to measure the oscillatory activity in theta band and synchrony between the principal executive areas with excellent temporal resolution. The effects of alcohol on theta activity and co-oscillations between the ACC and the latPFC are examined as a function of conflict elicited by the Stroop interference task. We hypothesize that increased cognitive demands are associated with greater functional synchrony and that alcohol-induced dysregulation of synchronous activity of the medial and lateral prefrontal cortices underlies impairments in cognitive control.

протокол

This experimental protocol has been approved by the Human Subjects Protection Committee at the University of California, San Diego. 1. Human Subjects Recruit healthy right-handed adult volunteers, obtain their consent, and screen them on the inclusion/exclusion criteria. NOTE: In this study, twenty young, healthy individuals (mean ± standard deviation [SD] age = 25.3 ± 4.4 years) including 8 women were recruited who drink in moderation, who have never been in treatment or arrested for drug or alcohol related offenses, who report no alcoholism-related symptoms on the Short Michigan Alcoholism Screening Test28, who do not smoke nor use illegal substances, who do not have a history of neuropsychiatric disorders or any current health problems, and who are medication free and have no internal ferromagnetic objects or implants. 2. Experimental Design Scan each participant four times, including three MEG sessions (a no-beverage introductory session and two experimental beverage sessions in which alcohol and placebo are administered in a counterbalanced manner), and one structural MRI scan. NOTE: In this within-subject design, participants serve as their own controls by participating in both alcohol and placebo sessions. This design reduces error variance and increases statistical power by minimizing influence of individual variability in brain anatomy, activity patterns, and alcohol metabolism. 3. Collecting MEG Scans Perform familiarization session. During the initial introductory session, administer questionnaires to obtain more information about the participants' medical history, their drinking patterns and severity of alcoholism-related symptoms28,29, family history of alcoholism30, and personality traits including impulsivity31,32. Carry out an initial recording in the MEG scanner following the protocol described below in steps 3.2, 3.3, and 3.5. Do not provide any beverage. Explain the task and run the practice version allowing participants to get familiarized with it beforehand. NOTE: The acclimation to the experimental situation serves the purpose of minimizing potential effects of situation-induced arousal33, thereby equating subsequent alcohol and placebo sessions on that dimension. Perform the alcohol/placebo experimental sessions. NOTE: Follow the same experimental procedures during both alcohol and placebo sessions with the exception of the administered beverage. Counterbalance beverage order by administering alcohol beverage first to one half of participants and placebo to the other half in a random order. Upon their arrival to the MEG lab, run a brief test scan by putting the participant in the scanner and checking the channels for possible magnetization. Measure their weight. Screen them with an electronic breathalyzer. Query them about compliance with the requirements to abstain from alcohol for 48 h and from food for 3 h prior to the experiment. Collect urine samples for a multi-drug test panel from all participants and exclude those who test positive for any drug. In addition, check female participants for pregnancy with a urine test and exclude those who test positive or if they suspect that they might be pregnant. Assess dynamic changes in the subjective effects of alcohol by asking participants to rate their momentary feelings and states on a standardized scale34 prior to drinking and on two additional occasions during the experiment – on the ascending limb (~15 min after consuming beverage) and descending limb of the breath alcohol concentration curve (BrAC), after the MEG recording. Administer a practice run of the Stroop task on a laptop with stimulus presentation software to ensure that the participants understand the task before recording. NOTE: This version of the Stroop task combines reading and color naming (Figure 1). The congruent condition consists of color words (i.e., red, green, blue, yellow) that are printed in the matching font color (i.e., the word "green" is printed in green). In the incongruous condition, color words are printed in color that does not match their meaning (i.e., the word "green" is printed in yellow). Ask the participants to press one of four buttons corresponding to the font color whenever a word is written in color, or, when a word is written in gray, to press a button corresponding to the meaning of the word18,23. Prepare the MEG/EEG recording. NOTE: Details of MEG data acquisition have been described in previous publications35,36,37. Position the EEG cap or individual EEG electrodes on the head of the participant and check that all impedances are below 5 kΩ. Attach the head position indicator (HPI) coils on either side of the forehead and behind each ear. NOTE: This step is specific to Neuromag systems. Digitize positions of the fiducial points including the nasion and two preauricular points, positions of HPI coils, EEG electrodes, and obtain a large number of additional points (~200) delineating the head shape. Use this information for the co-registration with anatomical MRI images (Figure 2). Administer beverage. Prepare alcohol beverage by mixing premium quality vodka with chilled orange juice (25% v/v), based on each participant's gender and weight (0.60 g/kg alcohol for men, 0.55 g/kg alcohol for women), targeting a BrAC of 0.06. Serve the same volume of orange juice in glasses with rims swabbed with vodka as a placebo beverage. Ask the participant to consume the beverage in approximately 10 min. Check the participants' BrAC with the breathalyzer starting at ~15 min after drinking and then every 5 min until they enter the recording chamber. Since electronic devices cannot be used in the shielded room, use a saliva alcohol test, which consists of a cotton swab that is saturated in saliva and is inserted into a receptacle that provides a readout. Acquire MEG/EEG data. Position the participant comfortably in the scanner. Since the prefrontal activity is of particular interest, ensure that the participant is positioned so that his/her head is touching the top of the helmet and is aligned along the front. NOTE: Head position can affect activity estimates in significant ways because the magnetic field gradients decrease with the cube of the distance between the sensors and the brain sources39. Connect HPI coils and all of the electrodes to their respective inputs on the scanner. Position response pads so that the buttons can be pressed comfortably. Ascertain that the font is clearly legible on the projection screen in front of the participant. Back in the console room, check that the intercom is functioning properly. Remind the participant to minimize blinking and to avoid movements including head motion caused by talking. Instruct the participant to reply to questions by pressing response buttons instead. Check that all response and stimulus triggers are recorded correctly. Examine all channels for artifacts and measure the head position in the scanner. Start data acquisition and begin the task. Give breaks every ~2.5 min to rest the eyes. Save the data upon task completion and escort the participant out of the recording chamber. When the participant has exited the scanner, acquire approximately two minutes of data from the empty room as a measure of instrumental noise. Ask the participant to rate perceived task difficulty, content of the imbibed beverage, how intoxicated they felt, as well as their momentary moods and feelings34. 4. Image Acquisition and Cortical Reconstruction of Structural MRI Obtain a high-resolution anatomical MRI scan for each participant, and reconstruct each participant's cortical surface with FreeSurfer software40,41,42. Use the inner skull surface derived from the segmented structural MRI images to generate a boundary element model of the volume conductor, which is used to provide a model for the forward solution that is consistent with each individual's brain anatomy43,44. 5. MEG Data Analysis NOTE: Analyze the data with the anatomically-constrained MEG approach which uses each participant's reconstructed cortical surface to constrain source estimates to the cortical ribbon40,45,46. The analysis stream relies on custom functions with dependencies on publicly available packages including FieldTrip47, EEGLab48, and MNE49. During data preprocessing, use a permissive band-pass filter (e.g., 0.1 – 100 Hz) and epoch data with respect to stimulus onset into segments that include padding intervals on each end (e.g., -600 to 1100 ms for an interval of interest spanning -300 to 800 ms after the removal of padding). Remove noisy and flat channels, as well as trials containing artifacts by visual inspection and using threshold-based rejection. Use independent component analysis48 to remove eyeblink and heartbeat artifacts. Eliminate trials with incorrect responses. Apply Morlet wavelets (Figure 3)47 to calculate complex power spectrum for each trial in 1 Hz increments for theta frequency band (4 – 7 Hz). Remove any additional artifacts. Compute the noise covariance from empty room data. Co-register the MEG data with MRI images using the three-dimensional (3D) head digitization information (Figure 2). Open the MRIlab module. Select File| Open| Select subject's structural MRI. Select File| Import| Isotrak data| select raw data.fif file| Make Points. Select Windows| Landmarks| Adjust fiducial landmarks until co-registration of MEG data and MRI are acceptable. Select File| Save. Calculate noise-sensitivity normalized estimates of theta source power and phase with a spectral dynamic statistical mapping approach18,50. Express event-related theta source power as percent signal change relative to baseline. Create group averages of event-related theta source power by morphing each participant's estimates onto an average cortical representation51. Visualize the source estimates on an inflated average surface to enhance visibility of sulcal estimates (Figure 4). Open the MNE software. Select File| Load Surface| Load inflated group-average FreeSurfer cortical surface. Select File| Manage overlays| Load stc| Load group-averaged data| Select loaded file from available overlays. Select overlay type as Other. Adjust Color Scale thresholding| Show. View brain movies and examine spatio-temporal stages of processing by identifying areas and time windows characterized by highest activation. Create unbiased regions of interest (ROIs) based on overall group-averaged estimates to incorporate cortical locations with most notable source power. Calculate time courses for each subject, condition, and ROI (Figure 5). Submit the obtained theta source power estimates to the statistical analysis. Extract time windows of interest from each ROI time course and perform analysis of variance (ANOVA) with beverage (alcohol, placebo) and trial type (congruous, incongruous) as within subject factors. Use a nonparametric cluster-based permutation test52 to examine beverage and condition comparisons of event-related theta power as well as phase-locking values (PLV). Estimate task-related changes in the long-range synchronization between the main activation foci in the ACC and the latPFC by computing the PLV12. Express PLV as percent change relative to baseline. NOTE: The PLV is an indicator of consistency of the phase angle between the two ROIs across trials as it measures the extent to which they co-oscillate at a particular frequency and in real time (Movie 1). Calculate correlations between ROI MEG activity estimates, indices of behavioral performance, and questionnaire scores to inform interpretation of the observed results.

Representative Results

Behavioral results indicate that the Stroop task successfully manipulated response interference because the accuracy was the lowest and the response times the longest on incongruous trials (Figure 6). Alcohol intoxication lowered accuracy but did not affect reaction times18. The spatio-temporal sequence of activity in theta frequency band revealed with the aMEG approach is ov…

Discussion

The multimodal imaging method used in this study comprises distributed source modeling of the temporally precise MEG signal along with spatial constraints of inverse estimates derived from each participant's structural MRI. The aMEG approach combines the strengths of these techniques to provide insight into the spatio-temporal stages of oscillatory dynamics and the long-range integration subserving cognitive control. This method provides greater temporal precision than other neuroimaging techniques such as fMRI-BOLD …

Materials

Elekta Neuromag	Elekta		Magnetoencephalography system
1.5 T GE EXCITE HG	General Electric		Magnetic Resonance Imaging scanner
Gold Cup Electrodes	OpenBCI		Electroencephalography electrodes for optional simultaneous EEG recording
Prep Check Impedance Meter	General Devices		Check electrode impedances
HPI Coils	Elekta		Head position indicator coils for co-registration
Alcotest	Draeger		Breathalyzer
Fiber Optic Response Pad	Current Designs, Inc		MEG-compatible response pad
Grey Goose Vodka	Bacardi		Vodka is used during the alcohol session
Orange Juice	Naked		Orange juice is used as the beverage during the placebo session as well as mixed with vodka during the alcohol session
Discover Drug Test Card	American Screening Corp		Multi-screen drug test
QED Saliva Alcohol Test	OraSure Technologies		Saliva alcohol test
Urine Hcg Test Strips	Joylive		Pregnancy test
Short Michigan Alcohol Screening Test	Selzer et al., 1975		Alcoholism screening questionnaire
Zuckerman Sensation Seeking Scale	Zuckerman, 1971		Questionnaire: disinhibitory, novelty-seeking, and socialization traits
Eysenck Impulsivity Inventory	Eysenck & Eysenck, 1978		Questionnaire: impulsivity traits
Eysenck Personality Questionnaire	Eysenck & Eysenck, 1975		Questionnaire: personality traits
Biphasic Alcohol Effects Scale	Martin et al., 1993		Questionnaire: subjective experience of the effects of alcohol