Abstract

Real-time fMRI (rtfMRI) neurofeedback (NF) facilitates volitional control over brain activity and the modulation of associated mental functions. The NF signals of traditional rtfMRI-NF studies predominantly reflect neuronal activity within ROIs. In this study, we describe a novel rtfMRI-NF approach that includes a functional connectivity (FC) component in the NF signal (FC-added rtfMRI-NF). We estimated the efficacy of the FC-added rtfMRI-NF method by applying it to nicotine-dependent heavy smokers in an effort to reduce cigarette craving. ACC and medial pFC as well as the posterior cingulate cortex and precuneus are associated with cigarette craving and were chosen as ROIs. Fourteen heavy smokers were randomly assigned to receive one of two types of NF: traditional activity-based rtfMRI-NF or FC-added rtfMRI-NF. Participants received rtfMRI-NF training during two separate visits after overnight smoking cessation, and cigarette craving score was assessed. The FC-added rtfMRI-NF resulted in greater neuronal activity and increased FC between the targeted ROIs than the traditional activity-based rtfMRI-NF and resulted in lower craving score. In the FC-added rtfMRI-NF condition, the average of neuronal activity and FC was tightly associated with craving score (Bonferroni-corrected p = .028). However, in the activity-based rtfMRI-NF condition, no association was detected (uncorrected p > .081). Non-rtfMRI data analysis also showed enhanced neuronal activity and FC with FC-added NF than with activity-based NF. These results demonstrate that FC-added rtfMRI-NF facilitates greater volitional control over brain activity and connectivity and greater modulation of mental function than activity-based rtfMRI-NF.

INTRODUCTION

A series of studies have provided evidence for the feasibility of voluntary modulation of regional neuronal activity in the human brain using (near) real-time fMRI (rtfMRI)-based neurofeedback (NF) approaches that incorporate BOLD-based neuronal activity information (Arns, Heinrich, & Strehl, 2014; Garrison et al., 2013; Soldati, Calhoun, Bruzzone, & Jovicich, 2013; Caria, Sitaram, & Birbaumer, 2012; Johnson et al., 2012; Linden et al., 2012; LaConte, 2011; Subramanian et al., 2011; Lee, Ryu, Jolesz, Cho, & Yoo, 2009; Lee, O'Leary, Park, Jolesz, & Yoo, 2008; Yoo, Lee, O'Leary, Panych, & Jolesz, 2008; Weiskopf et al., 2007; Yoo et al., 2007; deCharms et al., 2005; Weiskopf et al., 2004; Posse et al., 2003; Yoo & Jolesz, 2002; see review articles: Ruiz, Buyukturkoglu, Rana, Birbaumer, & Sitaram, 2014; Stoeckel et al., 2014; Sulzer et al., 2013; Weiskopf, 2012).

Interestingly, rtfMRI-NF methods, which are based on feedback of regional neuronal activity, have demonstrated not only the modulation of neuronal activity in targeted ROIs but also the modulation of neuronal connectivity patterns across multiple brain regions (Ruiz et al., 2014; Haller et al., 2013; Lee, Kim, & Yoo, 2012; Van De Ville et al., 2012; Rota, Handjaras, Sitaram, Birbaumer, & Dogil, 2011). These neuronal connectivity patterns include functional connectivity (FC) estimated using Pearson's correlation coefficient (CC; Haller et al., 2013; Rota et al., 2011) and effective connectivity patterns estimated using Granger causality analysis (Ruiz et al., 2011, 2014; Lee, Kim, & Yoo, 2012; Rota et al., 2011). As an example, Haller and colleagues (2013) conducted an rtfMRI-NF study to regulate neuronal activity within the auditory cortex and reported that the functionally connected networks within the auditory area were modulated by rtfMRI-NF.

Neuropsychiatric disorders such as schizophrenia and attention deficit hyperactivity disorder involve aberrant neuronal connectivity patterns across brain regions (Sulzer et al., 2013; Suskauer et al., 2008; Hoffman et al., 1995). For examples, Hoffman et al. (1995) reported reduced cortical connectivity in patients with schizophrenia because of a suppressed mesocorticolimbic dopaminergic system, and Suskauer et al. (2008) reported aberrant FC patterns between the rostral SMA and the anterior prefrontal areas in children with attention deficit hyperactivity disorder. These correlated or causative whole-brain neuronal connectivity patterns can be used in network- or connectivity-based NF. Network- or connectivity-based NF is likely to represent the sensorimotor and cognitive dysfunctions associated with neuropsychiatric disorders more accurately than region-based NF (Ruiz et al., 2014; Sulzer et al., 2013). However, efforts to use neuronal connectivity patterns as an NF signal and to modulate neuronal connectivity-based NF signals via rtfMRI-NF remain limited (Koush et al., 2013; Ruiz et al., 2011). In a previous attempt, Koush and colleagues (2013) proposed using effective connectivity as an explicit NF signal. In their study, participants were able to learn to covertly alternate their visuospatial attention according to a target cue with the help of effective connectivity information across the task-associated bilateral superior parietal lobule and visual areas (Koush et al., 2013).

To the best of our knowledge, no study has combined information on FC between ROIs and neuronal activity within ROIs as an NF signal for use within the rtfMRI-NF framework in an effort to modulate distinct FC patterns and associated mental function. In this study, we systematically investigated the efficacy of including FC information in NF signals at facilitating the modulation of FC patterns, as well as mental function, via an rtfMRI-NF method.

Specifically, we explored the efficacy of the proposed FC-added rtfMRI-NF method at facilitating real-time control of neuronal processes involved in resisting cigarette cravings. Brain regions implicated in cigarette cravings include ACC (Hartwell et al., 2011; Kober et al., 2010; Azizian et al., 2009; Brody et al., 2007; Smolka et al., 2006; Due, Huettel, Hall, & Rubin, 2002; Ernst et al., 2001), medial pFC (Sutherland, McHugh, Pariyadath, & Stein, 2012; Hartwell et al., 2011; Brody et al., 2007; Stein et al., 1998), OFC (Lee, Kim, & Kim, 2012; Hartwell et al., 2011; Smolka et al., 2006; Due et al., 2002), posterior cingulate cortex (Lee, Kim, & Kim, 2012; Hartwell et al., 2011; Azizian et al., 2009; Brody et al., 2007), and precuneus (Lee, Kim, & Kim, 2012; Hartwell et al., 2011; Brody et al., 2007). In addition, central reward-related regions such as the ventral striatum/nucleus accumbens, dorsal striatum, and amygdala are well known for their role in addiction, and the anterior insula is especially important in nicotine addiction (Sutherland et al., 2012; Naqvi, Rudrauf, Damasio, & Bechara, 2007). The rtfMRI-NF targeting the medial frontal area has successfully modulated neuronal activity in the ROIs involved in cigarette cravings (Canterberry et al., 2013; Hanlon et al., 2013; Li et al., 2013). However, the efficacy of traditional activity-based rtfMRI-NF at modulating neuronal activity and subjective cigarette cravings remains limited to light smokers, and this approach appears to require repeated rtfMRI-NF training sessions (Canterberry et al., 2013; Hanlon et al., 2013).

In this study, we aimed to investigate the efficacy of FC-added rtfMRI-NF in the modulation of brain activity related to cigarette cravings and, ultimately, the reduction of cravings in heavy smokers and after few rtfMRI-NF training sessions. We hypothesized that our proposed FC-added rtfMRI-NF method would result in greater modulation of neural activity and connectivity in brain areas implicated in cigarette cravings compared with a traditional activity-based rtfMRI-NF method.

METHODS

Participants

The institutional review board of Korea University approved the entire study protocol. Fourteen right-handed, male heavy smokers, without any current neurological or mental disorders, who were motivated to quit smoking but were not currently undergoing any treatment, provided written informed consent to participate in the study. The participants were randomly assigned to one of the two conditions: traditional, activity-based rtfMRI-NF or new, FC-added rtfMRI-NF. The participants of two conditions had similar sociodemographic characteristics and cognitive performance but a different age of onset of smoking (Table 1). The inclusion criteria were assessed the day before the fMRI scanning and were as follows: Fagerström Test of Nicotine Dependence score > 4 (Heatherton, Kozlowski, Frecker, & Fagerstorm, 1991), >5 years of smoking, and smoking >10 cigarettes per day. Exclusion criteria included the use of tobacco products other than cigarettes (e.g., nicotine patches/gum) and current use of a nicotine replacement therapy, such as bupropion, varenicline, or nortriptyline (Canterberry et al., 2013; Hanlon et al., 2013; Hartwell et al., 2011; Brody et al., 2007). The carbon monoxide (CO) level in exhaled breath was assessed using the piCO smokerlyzer (Bedfont Scientific, Ltd., Rochester, UK; mean ± SD = 21.2 ± 4.1 ppm) to confirm that the participants were heavy smokers (Heffner et al., 2013; Selby et al., 2013; Cornelius, Lynch, Martin, Cornelius, & Clark, 2001). Each of the 14 participants (seven participants for each NF condition) was enrolled in two visits spaced 1 week apart. Before each visit, they were asked to abstain from smoking cigarettes for at least 6 hr.

Table 1. 

Demographic and Cognitive Performance Information of the Participants Obtained during a Face-to-Face Interview, with Groups Stratified according to the Two NF Conditions

Activity-based NF (n = 7)FC-added NF (n = 7)p Value (between NF Conditions)
Age (years) 26.00 (2.16) 26.00 (1.29) p ≈ 1.00 
Handednessa 83.97 (8.68) 85.00 (8.68) p = .83 
Education (years) 20.00 (2.15) 20.00 (1.29) p ≈ 1.00 
WRAT-3 reading score 92.04 (3.18) 92.05 (5.01) p ≈ 1.00 
Digit n-back test 
 1-back (%) 93.71 (8.79) 98.71 (1.98) p = .17 
 2-back (%) 92.29 (7.18) 94.14 (2.12) p = .52 
 3-back (%) 87.83 (8.18) 85.86 (9.30) p = .70 
 4-back (%) 81.71 (10.97) 83.43 (12.38) p = .79 
 5-back (%) 81.14 (11.64) 80.57 (10.37) p = .92 
FTND 4.71 (0.95) 5.00 (1.00) p = .59 
Years of cigarette smoking 8.00 (2.24) 6.71 (1.38) p = .22 
Cigarettes/day 17.29 (2.29) 15.71 (3.45) p = .34 
CO level 20.14 (4.30) 22.29 (3.99) p = .35 
Pack years smoked 6.92 (2.21) 5.29 (1.80) p = .15 
Age of onset of smoking 19.00 (1.00) 20.29 (0.488) p = .01 
Activity-based NF (n = 7)FC-added NF (n = 7)p Value (between NF Conditions)
Age (years) 26.00 (2.16) 26.00 (1.29) p ≈ 1.00 
Handednessa 83.97 (8.68) 85.00 (8.68) p = .83 
Education (years) 20.00 (2.15) 20.00 (1.29) p ≈ 1.00 
WRAT-3 reading score 92.04 (3.18) 92.05 (5.01) p ≈ 1.00 
Digit n-back test 
 1-back (%) 93.71 (8.79) 98.71 (1.98) p = .17 
 2-back (%) 92.29 (7.18) 94.14 (2.12) p = .52 
 3-back (%) 87.83 (8.18) 85.86 (9.30) p = .70 
 4-back (%) 81.71 (10.97) 83.43 (12.38) p = .79 
 5-back (%) 81.14 (11.64) 80.57 (10.37) p = .92 
FTND 4.71 (0.95) 5.00 (1.00) p = .59 
Years of cigarette smoking 8.00 (2.24) 6.71 (1.38) p = .22 
Cigarettes/day 17.29 (2.29) 15.71 (3.45) p = .34 
CO level 20.14 (4.30) 22.29 (3.99) p = .35 
Pack years smoked 6.92 (2.21) 5.29 (1.80) p = .15 
Age of onset of smoking 19.00 (1.00) 20.29 (0.488) p = .01 

The data are presented as the mean values (and standard deviation); p values were obtained from two-sample t tests between the two NF conditions.

WRAT = wide range achievement verbal fluency test (Wilkinson & Robertson, 2006); FTND = Fagerström Test of Nicotine Dependence (Heatherton et al., 1991); CO = carbon monoxide.

aHandedness is measured from Edinburg's Handedness Inventory (Oldfield, 1971).

Experimental Setup

As illustrated in Figure 1A, each participant underwent six rtfMRI-NF runs during each of the two visits. Each rtfMRI-NF run lasted 258 sec and consisted of the following: (1) a calibration period to align the coordinates of an MR-compatible eye tracker (NordicNeuroLab, Bergen, Norway; www.nordicneurolab.com), (2) a period of fixation to a white cross on a black screen, (3) presentation of the “ready” command, (4) presentation of a video clip showing smoking (i.e., the rtfMRI-NF period), (5) a period for subjective ratings of cigarette cravings, and (6) a fixation period at the end of the scan. The NF information (i.e., either neuronal activity or the mean of neuronal activity and FC) was delivered to participants by varying the opacity of the smoking scenes in the smoking video clips via alpha compositing (Hanson, Gagliardi, & Hanson, 2009), as previously described (Cerf et al., 2010). Video stimuli elicit stronger cigarette cravings than image stimuli (Sargent, Morgenstern, Isensee, & Hanewinkel, 2009) and subsequently elicit greater cigarette consumption (Shmueli, Prochaska, & Glantz, 2010). An opacity change in a video stimulus that is determined according to cognitive performance provides an efficient interface for participants to effectively regulate their cognitive function (Pessaux et al., 2014; Beauchamp, Pasalar, & Ro, 2010; Livingston et al., 2003). Livingston et al. (2003) investigated the efficacy of opacity change of video as a tool to provide an augmented reality environment and to increase an accuracy of participants to detect a target location. Thus, the delivery of NF signals to the participants was designed to be more intuitive, more appealing, and more straightforward than the frequently used interface of graphs (e.g., lines, bars, and thermometers/odometers; Canterberry et al., 2013; Hanlon et al., 2013; Li et al., 2013; LaConte, 2011; Goebel, Zilverstand, & Sorger, 2010; Yoo et al., 2007, 2008; Weiskopf et al., 2004).

Figure 1. 

Experimental setup of rtfMRI-NF. (A) Flow diagram showing the protocol and task paradigms for each visit (top) and each rtfMRI-NF run (six runs per visit; bottom). (B) Detailed information on the specific phase of each rtfMRI-NF run. (C) Steps involved in extracting the NF signal and delivering the NF signal to participants via the opacity of the smoking scene. NA = neuronal activity.

Figure 1. 

Experimental setup of rtfMRI-NF. (A) Flow diagram showing the protocol and task paradigms for each visit (top) and each rtfMRI-NF run (six runs per visit; bottom). (B) Detailed information on the specific phase of each rtfMRI-NF run. (C) Steps involved in extracting the NF signal and delivering the NF signal to participants via the opacity of the smoking scene. NA = neuronal activity.

Twelve video clips (each of 3-min duration) containing male smokers lighting and smoking a cigarette were collected via an Internet search. The order of presentation of the 12 video clips was counterbalanced across the 12 rtfMRI-NF runs for each participant and across the two NF conditions to minimize potential confounding effects attributable to stimulus order (Huettel, Song, & McCarthy, 2004). While the participants watched the video clips, they were asked to deploy any mental strategy that they chose to resist the urge to smoke. Participants were not given explicit strategies to use, were not trained in a list of potential strategies to use before NF, and were asked to report the mental strategies that they used after the rtfMRI-NF experiment. The instruction given to the participant was “keep looking at the video and try to resist the urge to smoke using a covert strategy.” The mental strategies adopted by the participants were compared between the two NF conditions using Pearson's chi-squared test for categorical variables (Rance, Ruttorf, Nees, Schad, & Flor, 2014; Izuma, Saito, & Sadato, 2010; Schiff et al., 2007; Dinn, Aycicegi, & Harris, 2004; Colvin, Dunbar, & Grafman, 2001) performed in SPSS Statistics software (version 21; IBM Corporation, Armonk, NY). In this test, the null hypothesis was that the mental strategies were similar across the two NF conditions (i.e., the number of participants who employed each mental strategy was the same across the two NF conditions).

The participants were allowed to have a short rest period during an rtfMRI-NF run. The opacity of the video clip was continuously controlled based on the level of the NF signal, such that the image was dark (opaque) when the NF signal level was at a maximum (i.e., 1) and the image was not altered (transparent) when the NF signal level was at a minimum (i.e., 0), as shown in Figures 2A and 2B. The participants were informed that the NF signal reflected in the opacity of the smoking scene was delayed by approximately 5–6 sec because of the hemodynamic response (Huettel et al., 2004).

Figure 2. 

Illustration of the way in which the PSC TS within ROIs1 (cyan line) and the resulting NF signal (purple line) were reflected in the opacity of the smoking scenes delivered in real time for a randomly selected participant in the activity-based NF condition (A). Illustration of the way in which the TS of the average of neuronal activity within ROIs1 and FC between ROIs1 and ROIs2 (cyan line) and the resulting NF signal (purple line) were reflected in the opacity of the smoking scenes delivered in real time for a randomly selected participant in the FC-added NF condition (B). The BOLD intensities corresponding to values less than the 10th percentile (purple dots) and greater than the 90th percentile (orange dots) shown on the averaged BOLD TS used in the off-line analysis (C). NA = neuronal activity; TS = time series.

Figure 2. 

Illustration of the way in which the PSC TS within ROIs1 (cyan line) and the resulting NF signal (purple line) were reflected in the opacity of the smoking scenes delivered in real time for a randomly selected participant in the activity-based NF condition (A). Illustration of the way in which the TS of the average of neuronal activity within ROIs1 and FC between ROIs1 and ROIs2 (cyan line) and the resulting NF signal (purple line) were reflected in the opacity of the smoking scenes delivered in real time for a randomly selected participant in the FC-added NF condition (B). The BOLD intensities corresponding to values less than the 10th percentile (purple dots) and greater than the 90th percentile (orange dots) shown on the averaged BOLD TS used in the off-line analysis (C). NA = neuronal activity; TS = time series.

An MR-compatible eye tracker was used to monitor whether the participants remained alert throughout the experiments. Immediately after the rtfMRI-NF period, the participants reported their current craving score (min/max = 1/10) by pressing a button on a fiber-optic response pad (Current Design, Philadelphia, PA; www.curdes.com), as visually guided on the screen. Additional behavioral scores, including a performance score (i.e., the participants' perception of their success in changing the video opacity and thus modulating their brain activity, min/max = 1/10) and a stimulus strength score (i.e., how much the smoking video clip elicited a desire to smoke, min/max = 1/10), were reported verbally by the participants after each run. The craving score, performance score, and stimulus strength score were compared across the two NF conditions using repeated-measures MANOVA (Lewis, Talkington, Puce, Engel, & Frum, 2011; Ojima, Nakamura, Matsuba-Kurita, Hoshino, & Hagiwara, 2011) of Visit (first and second; within-subject factor) and Condition (activity-based NF and FC-added NF; between-subject factor) with Abstained hours and CO level included as covarying factors.

At each visit, two separate fMRI sessions were administered in a non-real-time fashion, before and after the six rtfMRI-NF runs (noted as “pre-rtfMRI” and “post-rtfMRI,” respectively). These non-rtfMRI sessions were used to evaluate the efficacy of the rtfMRI-NF by probing the before-and-after effects from the NF training. During these pre-rtfMRI and post-rtfMRI runs, participants were asked to view five blocks of smoking images, five blocks of nonsmoking images, and 10 blocks of abstract fixation images with same magnitude responses to paired smoking/nonsmoking images, but with random phases (Holsen et al., 2012; Lee, Kim, & Kim, 2012). In each of the three types of image blocks (smoking, nonsmoking, and fixation), five images were presented over a period of 15 sec (i.e., 3 sec per image). During the five smoking image blocks, participants were asked to resist the urge to smoke using the same strategy adopted in the rtfMRI-NF runs. Immediately after each pre-rtfMRI or post-rtfMRI run, participants rated their current craving score.

Imaging Parameters

BOLD fMRI data were acquired using a 3-T Siemens Tim-Trio scanner with a 12-channel head coil (Erlangen, Germany). A standard gradient-echo EPI pulse sequence was applied to measure the BOLD intensity associated with neuronal activity across whole-brain areas (Huettel et al., 2004). The EPI parameters of the rtfMRI-NF run were as follows: repetition time (TR) = 1000 msec, echo time = 24 msec, field of view = 24 × 24 cm2, matrix size = 64 × 64, voxel size = 3.75 × 3.75 × 7.0 mm3, flip angle = 90°, and 20 interleaved slices at approximately 30° oblique to the AC–PC line without a gap (Baumgartner, Knoch, Hotz, Eisenegger, & Fehr, 2011; Hampton, Bossaerts, & O'Doherty, 2008). The EPI parameters of the pre-rtfMRI and post-rtfMRI runs were slightly modified to reduce voxel size and were as follows: TR = 2000 msec and echo time = 30 msec, 36 interleaved slices at approximately 30° oblique to the AC–PC line without a gap, and voxel size = 3.75 × 3.75 × 4 mm3.

rtfMRI-NF Settings

Our rtfMRI-NF software was implemented in MATLAB (version 7.9.0 R2009b; Mathworks, Inc., Natick, MA) and was updated from our previously developed in-house codes (Lee, Kim, & Yoo, 2012; Lee et al., 2008, 2009; Yoo et al., 2007, 2008). The rtfMRI-NF software was installed on a laptop computer (Intel Core i5 2.4 GHz, 8-GB RAM, 256-GB SSD as hard drive, Windows 7). The laptop computer was connected to a console computer of the MRI scanner via a TCP/IP connection, such that the raw EPI volumes reconstructed in the MRI console computer were made available on the laptop computer in real time. During the rtfMRI-NF trials, six head motion parameters were estimated relative to the first EPI volume using the realignment implemented in SPM8 (www.fil.ion.ucl.ac.uk/spm), and the estimated head motions were subsequently corrected (Sitaram et al., 2011).

Identification of ROIs

On the basis of previous reports of the neuronal underpinnings of smoking addiction (Canterberry et al., 2013; Hanlon et al., 2013; Li et al., 2013; Hartwell et al., 2011; Azizian et al., 2009; Brody et al., 2007; Smolka et al., 2006; Due et al., 2002; Ernst et al., 2001), bilateral ACC, medial pFC, and OFC areas were defined as the anterior ROIs (ROIs1), and bilateral posterior cingulate cortex and precuneus areas were defined as the posterior ROIs (ROIs2; Table 2). The anterior part of the brain (i.e., ROIs1) has been widely reported as a cigarette-craving-related brain region (Lee, Kim, & Kim, 2012; Hartwell et al., 2011; Azizian et al., 2009; Brody et al., 2007; Due et al., 2002) and has been utilized in previous rtfMRI studies on cigarette craving/resistance (Canterberry et al., 2013; Hanlon et al., 2013; Li et al., 2013). However, the involvement of the posterior part of the brain (i.e., ROIs2) in cigarette craving is unclear (Hartwell et al., 2011; Brody et al., 2007). Therefore, we used ROIs1 rather than ROIs2 as the target region to extract neuronal activity for the activity-based rtfMRI-NF. The combination of neuronal activity in ROIs1 and FC between ROIs1 and ROIs2 was used as NF information for the FC-added rtfMRI-NF condition. During the calibration and fixation periods in each rtfMRI-NF run (45 sec; Figure 1), the first EPI volume was warped directly to the Montreal Neurological Institute (MNI) space using spatial normalization implemented in SPM8. The automated anatomical labeling (AAL; Tzourio-Mazoyer et al., 2002) and Brodmann's area (BA) templates (Brodmann, 1909) in the corresponding space (Lee et al., 2008) were therefore available for the identification of ROIs. From the registered anatomical information of the EPI volume of each participant, the voxels corresponding to ROIs1 and ROIs2 were identified.

Table 2. 

Definitions of ROIs Using the AAL and BA Templates

ROIs1ROIs2
AAL Superior frontal cortex (L/R), medial OFC (L/R) Posterior cingulate cortex (L/R) 
ACC (L/R) Precuneus (L/R) 
BA 10, 11, 12, 24, 32, 33 23, 26, 29, 30, 31 
ROIs1ROIs2
AAL Superior frontal cortex (L/R), medial OFC (L/R) Posterior cingulate cortex (L/R) 
ACC (L/R) Precuneus (L/R) 
BA 10, 11, 12, 24, 32, 33 23, 26, 29, 30, 31 

The regions defined by the AAL and BA templates were intersected and used as each of the two ROIs.

L = left; R = right; ROIs1 = ROIs in the anterior parts of the brain; ROIs2 = ROIs in the posterior parts of the brain.

Generation of NF Signal in Real Time

To calculate the NF information, the BOLD signal of each voxel within the ROIs was band-pass filtered from 0.008 to 0.1 Hz to remove artifacts including a low-frequency linear drift (Huettel et al., 2004) using the third-order elliptic digital filter via the ellip.m command in MATLAB. Then, the average BOLD intensity during the 25-sec cross-fixation period (i.e., 20–45 sec) was used as the baseline BOLD intensity, and the percentage BOLD signal change (PSC) was calculated relative to this baseline in a voxelwise manner. The PSC values during the latest three TR periods were averaged to reduce potential high-frequency fluctuations because of nonneuronal components such as cardiac- and respiratory-related fluctuations (Cauda et al., 2011; Vallesi, McIntosh, Shallice, & Stuss, 2009; Fox & Raichle, 2007). This temporal smoothing approach has been applied to reduce nonneuronal artifacts and to produce a reliable indication of neuronal activity, particularly in on-line analysis settings (Veit et al., 2012; Lee et al., 2009). For example, Veit and colleagues (2012) applied a moving window that covered the current TR (1.5-sec TR) and the preceding three TRs to temporally smooth the BOLD signal during on-line estimation of feedback information.

Unusually large or small PSCs can have potentially deleterious effects on the average PSC value across all voxels within an ROI, and the median PSC value across all voxels in ROIs1 at each time point (i.e., each EPI volume) was therefore adopted to represent the neuronal activity level of ROIs1. The median PSC from ROIs1 was used as the NF signal in the traditional activity-based method (Figure 2A). The FC between the BOLD signals (using all the time points before the current time point) of ROIs1 and ROIs2 was quantified using Pearson's CC. Then, the mean of this FC and the neuronal activity of ROIs1 was used as the NF signal in the FC-added method (Figure 2B). The processing time for each EPI volume to extract the NF signal was <1 sec. The PSC within ROIs1 was limited to a maximum of 1 (when the neuronal activity was equal to or above 1) and a minimum of −1 (when the neuronal activity was equal to or below −1), which are the same as the limits of the CC used to quantified FC between the two ROIs; thus, neuronal activity and FC contributed equally to the NF signal.

The smoking video clips were completely opaque (alpha compositing value of 1.0 set by controlling the image property in MATLAB) at an NF signal intensity of 1 and were transparent (alpha compositing value of 0) at an NF signal intensity ≤ 0, as exemplified in Figure 2A and B. Therefore, the participants learned to increase their NF signal to 1 to make the smoking scene darker. Participants were informed that the smoking scenes would be darkened if the degree of cigarette resistance was enhanced; thus, they were instructed to make the smoking scenes as dark as possible during the rtfMRI-NF period. Figure 2 shows the neuronal activity of ROIs1 (Figure 2A), the mean of the neuronal activity of ROIs1 and the FC between ROIs1 and ROIs2 (Figure 2B) acquired from a randomly selected run and participant, and the resulting NF signal (purple) along with the corresponding smoking video scenes.

Off-line Analysis: Preprocessing of rtfMRI-NF Data

The acquired EPI volumes from each rtfMRI-NF run were preprocessed using a series of steps, including head motion correction, spatial normalization, and spatial smoothing, using an 8-mm isotropic FWHM Gaussian kernel implemented in SPM8. During the rtfMRI-NF period, nonneuronal components, including physiological artifacts that were dominant within the white matter and cerebrospinal fluid, might have contaminated the BOLD signals. Therefore, we employed a PCA and least-squares-based artifact correction (Kim, Kim, & Lee, 2013; Chai, Castañón, Öngür, & Whitfield-Gabrieli, 2012; Lee, Kim, & Kim, 2012; Kim & Lee, 2011; Wager et al., 2009; Fransson & Marrelec, 2008; Behzadi, Restom, Liau, & Liu, 2007; Huettel et al., 2004; Lin et al., 2003). The resulting BOLD fMRI time-series data were further processed using temporal smoothing across three volume periods of 3-sec duration followed by temporal detrending to significantly reduce low-frequency linear drift noise (Lund, Madsen, Sidaros, Luo, & Nichols, 2006; Huettel et al., 2004; Smith et al., 1999).

Off-line Analysis: Calculation of Neuronal Activity and FC

The BOLD time-series data were averaged across all voxels in each of the two ROIs to calculate neuronal activity and FC. To estimate the PSC of neuronal activity in each of the two ROIs, the baseline BOLD intensity was defined as the average of data points below the 10th percentile of BOLD intensity throughout the rtfMRI-NF period (cf. average BOLD intensity during the 25-sec cross-fixation period as a baseline during on-line analysis). This baseline was assumed to reflect the short rests/pauses taken by the participant during the NF period. Then, the PSC intensity of the average of data points above the 90th percentile was defined as the neuronal activity that occurred because of an effort to resist cigarette cravings (cf. the average PSC intensity during the latest three TR time points as neuronal activity for each time point during on-line analysis). In addition, during the off-line analysis, any period of at least 3 sec (i.e., three TR periods or approximately 0.3 Hz) with sustained PSC intensity within the 3-dB range was automatically detected and used to calculate the average BOLD intensities below the 10th percentile and above the 90th percentile to minimize potential confounding artifacts resulting from fluctuations related to breathing motion (i.e., ∼0.3 Hz; Birn, Diamond, Smith, & Bandettini, 2006), as shown in Figure 2C.

To calculate FC, Pearson's CC between the average BOLD time series from the two ROIs was calculated using all time points during the rtfMRI-NF period and transformed into a z score using Fisher's r-to-z transformation (Leonardi et al., 2013; Kim & Lee, 2011; Fox, Zhang, Snyder, & Raichle, 2009; Zar, 1996). Note that, during on-line analysis, FC was calculated using all BOLD time-series data points before the current time point.

Neuronal activity and FC were compared between the two visits and between the two NF conditions using a mixed ANOVA with Visit as the within-subject factor and NF condition as the between-subject factor. As a post hoc test within each NF condition, a paired t test was used to compare neuronal activity and FC between the two visits. Furthermore, the neuronal activity, FC, and the mean of neuronal activity and FC obtained from the off-line analysis were compared with the results from the on-line analysis to evaluate the accuracy of the on-line method for calculating the NF signal. The results from the on-line and off-line analyses across the 12 rtfMRI-NF runs in two visits for each participant were compared using an orthogonal regression approach (Jobson, 1991). This orthogonal regression analysis is well suited when there are two dependent variables (Carter, Bowling, Reeck, & Huettel, 2012), and an implementation in the Minitab software toolbox was used (version 17; www.minitab.com).

Off-line Analysis: FC Patterns within the ROIs during rtfMRI-NF Trials

For ROIs1 and ROIs2, the average BOLD time series was calculated using the BOLD time series of all the voxels in the ROIs and was defined as the reference BOLD time series. Then, FC was calculated as Pearson's CC between the reference BOLD time series of ROIs1 or ROIs2 and the BOLD time series of each individual voxel in all areas across the two ROIs, followed by Fisher's r-to-z transformation. The resulting z scored FC patterns were compared across the two NF conditions in a voxelwise manner using a two-sample t test. This was performed separately for the first and second visits. The voxels that had significantly different FC between the two NF conditions (p < 10−3 with a minimum of 10 voxels) were identified for each visit and for each reference BOLD time series.

Off-line Analysis: Regression Analysis Using Craving Score and Neuronal Activity and FC

For each participant, neuronal activity, FC, and craving score were averaged across the six rtfMRI runs obtained during each visit. Using these average values across all participants (n = 7 for each NF condition at each visit), an orthogonal regression analysis (Jobson, 1991) was conducted using the craving score and (a) the average neuronal activity in each of the two ROIs, (b) the average FC between the two ROIs, and (c) the mean of the average neuronal activity and average FC. Then, the Bonferroni correction was applied across the two independent regression tests (i.e., across two visits) for each NF condition for each outcome (neuronal activity, FC, and mean of neuronal activity and FC; Maysov & Kipyatkov, 2011; Mundfrom, Perrett, Schaffer, Piccone, & Roozeboom, 2006).

Evaluation of rtfMRI-NF Using Non-real-time Pre-rtfMRI and Post-rtfMRI Runs

The EPI volumes of each of the pre-rtfMRI and post-rtfMRI runs were preprocessed using the standard procedure of SPM8 in the order of realignment, slice timing correction, normalization, and spatial smoothing using an 8-mm isotropic FWHM Gaussian kernel. The preprocessed pre-rtfMRI/post-rtfMRI run was further preprocessed to remove the potential physiological artifacts estimated from white matter and cerebrospinal fluid, followed by temporal detrending, which was also adopted for the preprocessing of rtfMRI-NF data for off-line analysis.

The neuronal activity of each of the two ROIs and the FC between the two ROIs were estimated for the pre-rtfMRI and post-rtfMRI runs using the preprocessed BOLD time series, as described for the rtfMRI-NF runs. Specifically, to estimate neuronal activity, the BOLD time-series data were averaged across all voxels in each of the two ROIs, and the baseline BOLD intensity was defined as the average of the data points below the 10th percentile BOLD intensity during nonsmoking image blocks. Using this baseline, neuronal activity was defined as the PSC value of the average of the data points above the 90th percentile during smoking image blocks. FC was calculated as Pearson's CC between the average BOLD time series of the two ROIs and then transformed into a z score using Fisher's r-to-z transformation. A mixed ANOVA was performed to compare neuronal activity and FC between the two NF conditions (between-subject factor) and across the four non-rtfMRI runs (within-subject factor), and p values were Bonferroni corrected. In addition, an orthogonal regression analysis was performed to evaluate the relations between craving score and (a) neuronal activity, (b) FC, and (c) the mean of neuronal activity and FC. In this regression analysis, craving score, neuronal activity, FC, and the mean of neuronal activity and FC were averaged across the four non-rtfMRI runs for each participant. The p values from the regression analyses were adjusted using a Bonferroni correction across the two NF conditions for each outcome (neuronal activity, FC, and mean of neuronal activity and FC).

RESULTS

CO Level and Behavioral Data

Participants abstained from cigarette smoking for an average of 9.0 hr (SD = 1.4 hr). For each of the two conditions of participants (activity-based NF and FC-added NF), the CO level before each rtfMRI-NF visit was significantly lower from the CO level on the interview day (p < 10−4). CO level did not differ between the two NF conditions (seven participants per condition) for either of the two visits (p > .45) obtained from a two-sample t test. The mean (SD) number of voxels within ROIs1 and ROIs2 that were registered on-line during rtfMRI-NF runs was 203.7 (27.0) and 110.9 (26.3), respectively, across all of the 14 participants and 12 runs.

The mean (SD) craving score was slightly lower in the FC-added NF condition (4.95 [1.15] at the first visit and 4.55 [0.98] at the second visit) than that in the activity-based NF condition (5.26 [0.68] at the first visit and 5.36 [0.80] at the second visit), but this was not statistically significant (F(1, 12) = 3.17, p = .10 for main effect of NF condition). The performance score (i.e., the participants' perception of their success in changing the video opacity and thus modulating their brain activity) was similar in the two NF conditions (F(1, 7) = 0.056, p = .819) and the two visits (F(1, 7) = 0.273, p = .617). In both groups, the performance score was similar in the first and second visits (5.04 [0.09] in the first visit vs. 4.88 [1.03] in the second visit in the activity-based NF, p = .79 from a post hoc paired t test, and 4.83 [0.50] in the first visit vs. 5.30 [1.09] in the second visit in the FC-added NF, p = .18 from a post hoc paired t test). The stimulus strength score (i.e., how much the smoking video clip elicited a desire to smoke) was similar in the two NF conditions (F(1, 7) = 3.20, p = .117) and the two visits (F(1, 7) = 0.565, p = .477). Overall, there was no significant main effect of NF condition or visit for any variable except CO level (main effect of visit, p = .01; CO level was higher at the second visit than at the first visit) and no significant interaction between NF condition and visit for any variable. Table 3 summarizes the detailed information, along with the mental strategies used to resist cravings. There was no significant difference in the mental strategies adopted in the two NF conditions (χ2 = 4.20, uncorrected p = .241).

Table 3. 

Average (SD) of Abstained Hours, CO Levels, and Rated Scores from Participants across Two NF Conditions at Each Visit

MeasurementVisitActivity-based NF (n = 7)FC-added NF (n = 7)p Value between NF Conditionsp Value between Visits
Abstained hours First 8.86 (1.21) 9.43 (0.79) F(1, 12) = 0.613, p = .449 F(1, 12) = 0.814, p = .385 
Second 8.57 (2.37) 9.14 (1.07) [F(1, 7) = 2.247, p = .178] [F(1, 7) = 5.208, p = .056] 
CO level First 8.43 (3.55) 8.43 (2.99) F(1, 12) < 0.001, p = 1.000 F(1, 12) = 9.253, p = .010 
Second 9.57 (3.99) 9.57 (3.05) [F(1, 7) = 0.004, p = .950] [F(1, 7) = 5.660, p = .049] 
Craving score First 5.26 (0.68) 4.95 (1.15) F(1, 12) = 3.171, p = .100 F(1, 12) = 0.167, p = .690 
Second 5.36 (0.80) 4.55 (0.98) [F(1, 7) = 0.036, p = .854] [F(1, 7) = 1.144, p = .320] 
Stimulus strengtha First 5.29 (1.10) 6.27 (0.85)   
Second 5.17 (1.12) 5.90 (0.19) [F(1, 7) = 3.200, p = .117] [F(1, 7) = 0.565, p = .477] 
Performance scorea First 5.04 (0.85) 4.83 (0.50)   
Second 4.88 (1.03) 5.30 (1.09) [F(1, 7) = 0.056, p = .819] [F(1, 7) = 0.273, p = .617] 
 
Mental Strategies to Resist the Urge to Smoke (Number of Participants) Activity-based NF (n = 7) FC-added NF (n = 7) 
Recalling coursework assignments (n = 4) 
Memorizing English vocabulary (n = 2) 
Visualizing an adverse effect of smoking (n = 5) 
Passively viewing the smoking stimuli (n = 3) 
MeasurementVisitActivity-based NF (n = 7)FC-added NF (n = 7)p Value between NF Conditionsp Value between Visits
Abstained hours First 8.86 (1.21) 9.43 (0.79) F(1, 12) = 0.613, p = .449 F(1, 12) = 0.814, p = .385 
Second 8.57 (2.37) 9.14 (1.07) [F(1, 7) = 2.247, p = .178] [F(1, 7) = 5.208, p = .056] 
CO level First 8.43 (3.55) 8.43 (2.99) F(1, 12) < 0.001, p = 1.000 F(1, 12) = 9.253, p = .010 
Second 9.57 (3.99) 9.57 (3.05) [F(1, 7) = 0.004, p = .950] [F(1, 7) = 5.660, p = .049] 
Craving score First 5.26 (0.68) 4.95 (1.15) F(1, 12) = 3.171, p = .100 F(1, 12) = 0.167, p = .690 
Second 5.36 (0.80) 4.55 (0.98) [F(1, 7) = 0.036, p = .854] [F(1, 7) = 1.144, p = .320] 
Stimulus strengtha First 5.29 (1.10) 6.27 (0.85)   
Second 5.17 (1.12) 5.90 (0.19) [F(1, 7) = 3.200, p = .117] [F(1, 7) = 0.565, p = .477] 
Performance scorea First 5.04 (0.85) 4.83 (0.50)   
Second 4.88 (1.03) 5.30 (1.09) [F(1, 7) = 0.056, p = .819] [F(1, 7) = 0.273, p = .617] 
 
Mental Strategies to Resist the Urge to Smoke (Number of Participants) Activity-based NF (n = 7) FC-added NF (n = 7) 
Recalling coursework assignments (n = 4) 
Memorizing English vocabulary (n = 2) 
Visualizing an adverse effect of smoking (n = 5) 
Passively viewing the smoking stimuli (n = 3) 

The p values were obtained from repeated-measures MANOVA. Data are presented in mean (standard deviation).

[ ] = repeated MANOVA results using these nine samples; CO = carbon monoxide.

aA number of participants for each NF condition was less than 7 (i.e., four participants for activity-based NF and five participants for FC-added NF).

Neuronal Activity and Relation to Craving Score

The mean and standard error of the neuronal activity of ROIs1 across the seven participants in each NF condition at each visit is shown in Figure 3A. The neuronal activity in ROIs1 was similar across the two NF conditions during the first (p = .77) and second (p = .095) visits. However, in both conditions, neuronal activity increased significantly from the first visit to the second visit (activity-based NF: p = 1.0 × 10−5; FC-added NF: p = 1.8 × 10−8). Using a post hoc paired t test between two visits fixing an NF condition, the statistical significance of neuronal activity enhancement was stronger (i.e., p = 8.1 × 10−7 for the activity-based NF condition and p = 2.1 × 10−9 for the FC-added NF condition). In the FC-added NF condition, there was a significant negative association between neuronal activity in ROIs1 and craving score at the second visit (z = −8.28, corrected p = 2.42 × 10−8) and a marginally negative association between neuronal activity in ROIs1 and craving score at the first visit (z = −1.68, uncorrected p = .046). In the activity-based NF condition, there was a moderately negative association between neuronal activity of ROIs1 and craving score at the second visit (z = −1.95, uncorrected p = .026) but no association at the first visit (z = −0.22, uncorrected p = .413). As depicted in Figure 3B, neuronal activity in ROIs2 across the two visits was greater in the FC-added NF condition than that in the activity-based NF condition (p = 4.6 × 10−3). At the second visit, the statistical significance was enhanced (p = 3.3 × 10−4). In the FC-added NF condition, neuronal activity in ROIs2 was greater at the second visit than at the first visit (p = 1.2 × 10−5). Using a post hoc paired t test, a statistical significance between the two visits becomes substantially increased (i.e., p = 9.4 × 10−3 for the activity-based NF condition and p = 3.7 × 10−9 for the FC-added NF condition). There was no association between the neuronal activity in ROIs2 and craving score (uncorrected p > .069).

Figure 3. 

(Left) Bar plots of neuronal activity (NA) in ROIs1 (A) and ROIs2 (B) at each visit for each of the two NF conditions. (Right) The results of the regression analyses of NA and craving score at each visit in each NF condition. The p values are Bonferroni-corrected for multiple comparisons across two visits. A significant correlation (corrected p < .05) is indicated by boldface font and a thick black regression line. ROIs1 and ROIs2 are the ROIs in the anterior and posterior parts of the brain, respectively, as denoted in Table 2. Please refer to the Results section for the details. pcorrected = the Bonferroni-corrected p value.

Figure 3. 

(Left) Bar plots of neuronal activity (NA) in ROIs1 (A) and ROIs2 (B) at each visit for each of the two NF conditions. (Right) The results of the regression analyses of NA and craving score at each visit in each NF condition. The p values are Bonferroni-corrected for multiple comparisons across two visits. A significant correlation (corrected p < .05) is indicated by boldface font and a thick black regression line. ROIs1 and ROIs2 are the ROIs in the anterior and posterior parts of the brain, respectively, as denoted in Table 2. Please refer to the Results section for the details. pcorrected = the Bonferroni-corrected p value.

FC Patterns Modulated by rtfMRI-NF

Figure 4 shows the FC contrast between the two NF conditions during each of the two visits and in each of the two seed ROIs, highlighting the modulated FC spatial patterns. Overall, it is notable that, during each of the two visits, FC across the ROIs was stronger in the FC-added NF condition than that in the activity-based NF condition (p < 10−3 from two-sample t tests). There was virtually no enhancement of FC in the ROIs in the activity-based NF condition compared with the FC-added NF condition. Table 4 summarizes the detailed information.

Figure 4. 

Spatial patterns of FC attained with rtfMRI-NF. Data are a comparison of the two NF conditions at the first visit (left) and the second visit (right) with ROIs1 (top) and ROIs2 (bottom) used as seed region for the FC calculation. Please refer to the Methods and Results sections for details. ROIs1 and ROIs2 are the ROIs in the anterior and posterior parts of the brain, respectively, as denoted in Table 2. Table 4 summarizes the detailed information.

Figure 4. 

Spatial patterns of FC attained with rtfMRI-NF. Data are a comparison of the two NF conditions at the first visit (left) and the second visit (right) with ROIs1 (top) and ROIs2 (bottom) used as seed region for the FC calculation. Please refer to the Methods and Results sections for details. ROIs1 and ROIs2 are the ROIs in the anterior and posterior parts of the brain, respectively, as denoted in Table 2. Table 4 summarizes the detailed information.

Table 4. 

Identified Brain Regions with Active Voxels (Uncorrected p < 10−3 with a Minimum of 10 Voxels) from the Comparison between the Two NF Conditions Using the FC Spatial Patterns of Individual Participants (x, y, and z Are Expressed in the MNI Coordinates)

Seed ROIsContrastVisitBrain RegionsFocus in x, y, and z Axes (mm)SizePeak t Score
ROIs1 Activity-based NF > FC-added NF First  No suprathreshold cluster   
Second  No suprathreshold cluster   
FC-added NF > activity-based NF First mOFC, ACC, mSFG 0, 59, −2 243 7.18 
Precuneus, PCC −3, −43, 34 81 4.66 
Second Precuneus, PCC 6, −46, 13 32 5.66 
mOFC, ACC 0, 35, −8 82 5.50 
mOFC 6, 56, −2 14 4.96 
ROIs2 Activity-based NF > FC-added NF First PCC −12, −49, 25 25 5.09 
Second  No suprathreshold cluster   
FC-added NF > activity-based NF First mOFC, ACC 0, 47, −5 196 6.67 
PCC 3, −46, 22 38 5.44 
Second Precuneus, PCC 6, −46, 13 103 7.28 
Precuneus −15, −64, 28 18 5.82 
mOFC, ACC −3, 47, −5 32 4.26 
mOFC 6, −23, −11 11 3.86 
Seed ROIsContrastVisitBrain RegionsFocus in x, y, and z Axes (mm)SizePeak t Score
ROIs1 Activity-based NF > FC-added NF First  No suprathreshold cluster   
Second  No suprathreshold cluster   
FC-added NF > activity-based NF First mOFC, ACC, mSFG 0, 59, −2 243 7.18 
Precuneus, PCC −3, −43, 34 81 4.66 
Second Precuneus, PCC 6, −46, 13 32 5.66 
mOFC, ACC 0, 35, −8 82 5.50 
mOFC 6, 56, −2 14 4.96 
ROIs2 Activity-based NF > FC-added NF First PCC −12, −49, 25 25 5.09 
Second  No suprathreshold cluster   
FC-added NF > activity-based NF First mOFC, ACC 0, 47, −5 196 6.67 
PCC 3, −46, 22 38 5.44 
Second Precuneus, PCC 6, −46, 13 103 7.28 
Precuneus −15, −64, 28 18 5.82 
mOFC, ACC −3, 47, −5 32 4.26 
mOFC 6, −23, −11 11 3.86 

The ROIs1 and ROIs2 are the ROIs in the anterior and posterior parts of the brain, respectively, as denoted in Table 2.

mOFC = medial OFC; mSFG = medial superior frontal gyrus; PCC = posterior cingulate cortex.

FC between the Two ROIs and Association with Craving Score

Figure 5 shows the mean and standard error of the z scored FC between the two ROIs across the seven participants in each NF condition at each visit. Overall, FC was greater in the FC-added NF condition than in the activity-based NF condition (p = 5.3 × 10−4). The difference between the groups was greater at the second visit (p = 5.7 × 10−3) compared with the first visit (p = .029). In the FC-added NF condition, FC was greater at the second visit than at the first visit (p = .02 from a post hoc paired t test). FC was negatively associated with craving score at the second visit in the activity-based NF condition (z = −2.20, corrected p = .028) and in the FC-added NF condition (z = −1.67, uncorrected p = .047) but was not associated with craving score at the first visit in either condition (uncorrected p > .161).

Figure 5. 

FC between the two ROIs at each visit for each NF condition and the association between FC and craving score. The p values are Bonferroni-corrected for multiple comparisons across two visits. A significant correlation (corrected p < .05) is indicated by boldface font and a thick black regression line. Please refer to the Results section for the details. pcorrected = the Bonferroni-corrected p value.

Figure 5. 

FC between the two ROIs at each visit for each NF condition and the association between FC and craving score. The p values are Bonferroni-corrected for multiple comparisons across two visits. A significant correlation (corrected p < .05) is indicated by boldface font and a thick black regression line. Please refer to the Results section for the details. pcorrected = the Bonferroni-corrected p value.

Average of Neuronal Activity and FC and Its Association with Craving Score

Figure 6 represents the results of the orthogonal regression analyses across the seven participants in each NF group at each visit using (1) the mean of neuronal activity of ROIs1 (Figure 6A) or ROIs2 (Figure 6B) and FC and (2) craving score. In the FC-added NF condition, the mean of neuronal activity of ROIs1 and FC exhibited a significantly negative correlation with craving score at the second visit (z = −2.19, corrected p = .028). The mean of neuronal activity in ROIs2 and FC did not exhibit a correlation with craving score at either visit in either of the two NF conditions (uncorrected p > .081).

Figure 6. 

Association between the mean of neuronal activity (NA) and FC (i.e., (NA + FC) / 2) and craving score when NA is calculated from ROIs1 (A) and ROIs2 (B). The p values are Bonferroni-corrected for multiple comparisons across two visits. A significant correlation (corrected p < .05) is indicated by boldface font and a thick black regression line. ROIs1 and ROIs2 are the ROIs in the anterior and posterior parts of the brain, respectively, as denoted in Table 2. Please refer to the Results section for the details. pcorrected = the Bonferroni-corrected p value.

Figure 6. 

Association between the mean of neuronal activity (NA) and FC (i.e., (NA + FC) / 2) and craving score when NA is calculated from ROIs1 (A) and ROIs2 (B). The p values are Bonferroni-corrected for multiple comparisons across two visits. A significant correlation (corrected p < .05) is indicated by boldface font and a thick black regression line. ROIs1 and ROIs2 are the ROIs in the anterior and posterior parts of the brain, respectively, as denoted in Table 2. Please refer to the Results section for the details. pcorrected = the Bonferroni-corrected p value.

Figure 7 shows that the results of on-line and off-line analyses were highly correlated (uncorrected p < .05) in at least five of the seven participants (denoted as boldface fonts and thick regression lines) for each of the two NF conditions.

Figure 7. 

The neuronal activity (A, red), FC (B, blue), and mean of neuronal activity and FC (C, magenta) obtained from the on-line (horizontal axis) and off-line (vertical axis) analyses. A significant correlation (uncorrected p < .05) is indicated by boldface font and a thick black regression line. The column and row indicate the NF condition and the participant number in each NF condition, respectively. NA = neuronal activity.

Figure 7. 

The neuronal activity (A, red), FC (B, blue), and mean of neuronal activity and FC (C, magenta) obtained from the on-line (horizontal axis) and off-line (vertical axis) analyses. A significant correlation (uncorrected p < .05) is indicated by boldface font and a thick black regression line. The column and row indicate the NF condition and the participant number in each NF condition, respectively. NA = neuronal activity.

Neuronal Activity and FC from Pre-rtfMRI/Post-rtfMRI Data and Their Association with Craving Score

Figure 8 shows neuronal activity and FC, and their association with craving score, in each of the non-rtfMRI runs for each NF condition. The neuronal activity of both ROIs and the FC between ROIs in the pre-rtfMRI run of the first visit (i.e., a baseline condition) was similar in the two NF conditions (p > .05). There was a significant increase in neuronal activity over the four non-rtfMRI runs for ROIs1 (p = .013, main effect of run) and ROIs2 (p = .016, main effect of run). The simple main effects revealed that a significant increase was observed only in the FC-added NF condition (p = 1.4 × 10−3 for ROIs1 and p = 6.2 × 10−3 for ROIs2). The neuronal activity of ROIs1 in the post-rtfMRI run at the second visit was significantly higher from the FC-added NF than from the activity-based NF (p = .031 from a post hoc two-sample t test). At the second visit in the FC-added NF condition, the neuronal activity of ROIs1 was significantly higher in the post-rtfMRI run than in the pre-rtfMRI run (p = .029 from a post hoc paired t test). In both NF conditions, the neuronal activity in ROIs1 was higher in the post-rtfMRI at the second visit than in the pre-rtfMRI at the first visit (Bonferroni-corrected p = .012). In the FC-added NF condition, there was a marginally significant increase in FC across the four non-rtfMRI runs (p = .046, main effect of run).

Figure 8. 

Bar plots of neuronal activity (A), FC (B), and mean of neuronal activity and FC (C) within and/or between the ROIs in each of the non-real-time pre-rtfMRI and post-rtfMRI runs at the first and second visits. Scatter plots show the relation between the craving score and the neuronal activity (red), FC (blue), or mean of neuronal activity and FC (magenta) across the participants. The p values are Bonferroni-corrected for multiple comparisons across the two NF conditions. A significant correlation (corrected p < .05) is indicated by boldface font and a thick black regression line. Please refer to the Results section for the details. NA = neuronal activity; pcorrected = the Bonferroni-corrected p value.

Figure 8. 

Bar plots of neuronal activity (A), FC (B), and mean of neuronal activity and FC (C) within and/or between the ROIs in each of the non-real-time pre-rtfMRI and post-rtfMRI runs at the first and second visits. Scatter plots show the relation between the craving score and the neuronal activity (red), FC (blue), or mean of neuronal activity and FC (magenta) across the participants. The p values are Bonferroni-corrected for multiple comparisons across the two NF conditions. A significant correlation (corrected p < .05) is indicated by boldface font and a thick black regression line. Please refer to the Results section for the details. NA = neuronal activity; pcorrected = the Bonferroni-corrected p value.

The regression analysis revealed a significant correlation between neuronal activity level in each of the two ROIs and craving score in the FC-added NF condition (z = −2.07 and corrected p = .038 for ROIs1, z = −3.09 and corrected p = .002 for ROIs2) and between neuronal activity level in ROIs1 and craving score in the activity-based NF condition (z = −2.21, corrected p = .028). In the FC-added NF condition, there was a marginally negative association between FC and craving score (z = −1.67, uncorrected p = .046) and between the mean of neuronal activity and FC and craving score for ROIs1 (z = −1.88, uncorrected p = .030) and ROIs2 (z = −2.23, corrected p = .026).

DISCUSSION

Summary

In this study, we demonstrated the use of a novel FC-added rtfMRI-NF procedure to modulate neuronal activity and FC and subjective cigarette cravings. Compared with the traditional activity-based rtfMRI-NF method, our novel rtfMRI-NF method appeared to enhance neuronal activity in brain regions associated with cigarette craving resistance. Although neuronal activity in these regions was also enhanced in participants who received traditional activity-based rtfMRI-NF, the degree of association between neuronal activity and cigarette cravings was greater for the participants who received the FC-added rtfMRI-NF. The FC between the two ROIs was substantially higher in participants who received FC-added NF compared with those who received activity-based NF. Furthermore, the Bonferroni-corrected p value indicated that the mean of neuronal activity and FC was significantly associated with craving score only in participants who performed the FC-added NF. These results suggest that NF of FC between ROIs promoted the modulation of neuronal activity in relevant brain regions to a greater degree than NF of the neuronal activity level of the ROIs. There was a similar trend of neuronal activity and FC enhancement along with a tight link between neuronal activity within ROIs and FC between ROIs and craving score from the non-rtfMRI data analysis.

Associations between Neuronal Activity and Cigarette Cravings

According to an rtfMRI-NF study by Li and colleagues (2013), neuronal activity in ACC was positively correlated with the subjective craving rating when participants were asked to reduce cigarette craving, but neuronal activity in the middle frontal cortex was not correlated with craving score when the participants were asked to resist the urge to smoke. The results of our study indicate that neuronal activity in the anterior ROI (ROIs1), which included ACC and the medial OFC, and the posterior ROI (ROIs2), which included the posterior cingulate cortex and precuneus, was increased in both NF conditions, albeit to a greater extent with FC-added NF than with neuronal activity-based NF. In addition, the subjective craving score was associated with neuronal activity in the anterior ROIs but not with neuronal activity in the posterior ROIs. This is in line with the equivocal findings on the correlation between neuronal activity in posterior ROIs and cigarette craving in the literature (Hartwell et al., 2011; Brody et al., 2007).

In a study by Canterberry et al. (2013), neuronal activity in anterior ROIs did not correlate significantly with the severity of smoking addiction during the first two visits for rtfMRI-NF. However, by the third rtfMRI-NF visit, the severity of smoking addiction correlated with neuronal activity in ACC areas, and light smokers could more effectively reduce neuronal activity in this region while trying to resist cigarette cravings (Canterberry et al., 2013). By contrast, in our study, heavy smokers who received FC-added NF exhibited a significant association (1) between neuronal activity and cigarette craving score and (2) between the mean of neuronal activity and FC and cigarette craving score at the second rtfMRI-NF visit, according to the Bonferroni-corrected p value. Therefore, our FC-added NF training appears to facilitate modulation of neuronal activity, FC, and, ultimately, craving score in heavy smokers. However, we cannot completely rule out the possibility that differences between the two NF conditions in this study may have been influenced by the difference in the age of onset of smoking (Jackson, Sher, Cooper, & Wood, 2002; Taioli & Wynder, 1991).

Scaling of Neuronal Activity and Weighting of FC Information

In this study, values of neuronal activity above 1 or below −1 were limited to 1 and −1, respectively. This may have underestimated the performance of some participants because an increase in neuronal activity beyond 1 produced by the participant would not have had any effect on the NF signal. In addition, in the FC-added NF condition, we combined neuronal activity and FC using equal weighting for each of the two variables. Different ways of combining the neuronal activity and FC information could be used for FC-added NF, for example, by using different relative weighting of the two variables (Ruiz et al., 2011, 2014). Future studies should use differently weighted combinations of neuronal activity and FC information to investigate the role of FC information on rtfMRI-NF performance in a more systematic fashion.

Increased Neuronal Activity in the Two ROIs

Notably, neuronal activity in the posterior ROI was increased by traditional activity-based NF, despite the fact that feedback was only provided on neuronal activity in the anterior ROI. This might be because of inherent coactivation between anterior and posterior ROIs within so-called “default-mode networks” (DMNs; Kim & Lee, 2011; Fox et al., 2005). Assuming that the coactivation of the two ROIs observed in this study was because of intrinsic activity across DMN regions (Chai, Ofen, Gabrieli, & Whitfield-Gabrieli, 2014; Raichle & Snyder, 2007), it is possible that this intrinsic activity might be related to the cognitive functions associated with the response to, or interpretation of, visual cues associated with smoking. The participants reported reasonable task performance, as revealed by moderate performance scores, and they stayed alert throughout the runs, as reflected by their eye movements, which were monitored using an MR-compatible eye tracker. Therefore, we can exclude the possibility that the observed neuronal activity in the ROIs was attributable to inherent spontaneous activity. Instead, the intrinsic activity likely resulted from efforts to resist cigarette cravings during rtfMRI-NF training. Nonetheless, we cannot exclude the possibility that the enhanced neuronal activity and FC and the modulated craving score were because of efforts by the participants to ignore the smoking videos and to maintain a default-mode-like cognitive state.

The degree of modulation of FC was greater when the FC between ROIs1 and ROIs2 was included in the NF information. The two ROIs spatially overlapped with the DMNs of the resting-state FC patterns. Thus, these data suggest that the inclusion of the resting-state information (i.e., FC of DMNs) might have caused an early enhancement of FC between the two ROIs from the FC-added NF compared with the activity-based NF. Future studies that employ additional methods to monitor brain activity/connectivity, such as simultaneous electroencephalography and fMRI (Kim, Yoo, & Lee, 2015) with an additional acoustic oddball task (Czisch et al., 2012), are warranted to evaluate the fundamental mechanism by which incorporating FC information into the NF signal in rtfMRI-NF influences the modulation of neuronal activity and FC.

Dynamic Changes in FC

In this study, the FC between two ROIs during an rtfMRI run was calculated using all time points before the current time point. FC therefore was continuously updated and represented the overall level of FC within a single rtfMRI-NF run. In recent years, dynamic changes in FC have widely been investigated using a sliding temporal window approach, despite the potential susceptibility of this method to artifactual fluctuations from noise and the dependency on window size (Allen et al., 2014; Lindquist, Xu, Nebel, & Caffo, 2014; Zalesky, Fornito, Cocchi, Gollo, & Breakspear, 2014; Hutchison, Womelsdorf, Allen, et al., 2013; Hutchison, Womelsdorf, Gati, Everling, & Menon, 2013; Leonardi et al., 2013; Liu & Duyn, 2013). Using this approach, brain regions have been characterized depending on the degree of change in FC during an fMRI run (Zalesky et al., 2014; Hutchison, Womelsdorf, Allen, et al., 2013; Leonardi et al., 2013). For example, Zalesky and colleagues (2014) reported that the limbic areas showed relatively greater FC change (i.e., were more dynamic) than the visual areas, somatomotor areas, and DMNs, which showed relatively smaller FC change (i.e., were more static). The ROIs adopted in this study substantially overlap with the DMNs, and thus, our findings may not critically be altered if we used a sliding window approach to calculate FC. Use of a sliding window approach is warranted in future studies to enable dynamic changes in FC to be incorporated into the NF information.

Confounding Effects

Our off-line analysis included a physiological noise correction in addition to a motion correction and band-pass filtering that were conducted during on-line and off-line analyses. In most participants (i.e., at least five of the seven participants), the neuronal activity, FC, and the mean of neuronal activity and FC measured with on-line and off-line analyses were strongly correlated (shown in Figure 7; performed in each participant of the two NF conditions). This indicates that the NF information that the participants received during the rtfMRI-NF period was not critically confounded by physiological noise. Nonetheless, it is worth noting that our on-line preprocessing method during FC-added NF could be enhanced and supplemented by incorporating a physiological noise removal method such as the anatomical-component-based noise reduction method to minimize spurious correlations in FC (Chai et al., 2012; Behzadi et al., 2007). Using this anatomical-component-based noise reduction method, potential physiological noise in the BOLD time series within the white matter and cerebrospinal fluid areas can be estimated via principal components. Subsequently, a regression approach can reduce the physiological noise in the BOLD time series of gray matter and may thus increase the specificity of the FC quantified between the anterior and posterior parts of DMNs (Chai et al., 2012). The substantial increase in computational load required for this modification would need to be managed via a dedicated hardware system, such as the multicore processing capability of a graphic processing unit (gpgpu.org).

It is possible that the manipulation of the video clip also drove changes in unexpected brain activity and connectivity in addition to a brain process mediating the targeted cognitive functions (Veilleux, Conrad, & Kassel, 2013; Wagner, Dal Cin, Sargent, Kelley, & Heatherton, 2011; Beauchamp et al., 2010). For example, in previous cigarette craving studies, a smoking video cue has driven changes in the activity of an action observation network that includes the inferior parietal sulcus and inferior frontal gyrus, which are not readily activated when smoking images are used as visual cues (Wagner et al., 2011). However, it is important to note that any endogenous process that arose during manipulation of the smoking video clips in this study would likely have affected all participants across the two NF conditions. This common process used to manipulate the video stimuli across the two NF conditions can therefore be estimated using techniques such as canonical correlation analysis (Gaebler et al., 2014). By removing the common endogenous manipulation process, the brain process involved in resisting the urge to smoke can better be estimated, and thus, future study is warranted in this context.

Potential Limitations

It is important to note the limitations of our study. First, the sample size in each NF condition (n = 7) was relatively small, and a regression analysis using seven data points (i.e., seven participants) for each NF condition may be sensitive to single outlier point, although the Bonferroni correction that we applied might have reduced the likelihood of false positive results. However, this sample size is comparable with previous studies that have applied rtfMRI-NF methods to nicotine-use disorders (Canterberry et al., 2013; Hanlon et al., 2013; Li et al., 2013). Second, our ROIs were defined using anatomical atlases (i.e., AAL and BA) and based on the direct EPI to MNI warping, and inaccuracies inherent to this method could have resulted in neighboring voxels outside the ROI being included in the target ROI mask (Klein et al., 2009). In addition, the ROIs may include subregions with different functional characteristics. For example, ACC can be subdivided into subgenual, rostral, and dorsal parts, and each of these subregions appears to serve different functions (Torta & Cauda, 2011; Margulies et al., 2007). Thus, fine-tuning the ROIs based on the neuronal activation of each individual participant (i.e., functionally guided ROIs) estimated from analyses such as a general linear model and independent component analysis (Kim, Kim, & Lee, 2012) would increase the reliability of the NF information provided to participants (Sulzer et al., 2013; Veit et al., 2012; Hamilton, Glover, Hsu, Johnson, & Gotlib, 2011). In addition, a high-resolution structural MRI scan from each participant could be used to improve the accuracy of ROI definition using a standard normalization step or more sophisticated methods such as Diffeomorphic Anatomical Registration Through Exponentiated Lie Algebra, which were implemented in SPM8 (Klein et al., 2009; Ashburner, 2007). Furthermore, we did not assess intervention-related changes in smoking behavior or cravings outside the laboratory (West, Hajek, Stead, & Stapleton, 2005). The fact that the mental strategies used by the participants were not controlled during rtfMRI-NF runs in our study is a potential weakness, as the NF performance would be dependent on the adopted mental strategy (Sulzer et al., 2013; deCharms et al., 2005). Future investigation is warranted to evaluate the efficacy of our proposed method in a larger sample who were instructed to adopt a specific mental strategy investigated from a pilot study and with the use of additional smoking behavior and craving measures.

Future Studies

It is important to note that the neuronal activity and/or FC of the central reward regions, such as the ventral and dorsal striatum, nucleus accumbens, amygdala, and anterior insula, can be used as NF information in future studies (Sutherland et al., 2012; Naqvi et al., 2007). Beyond the application of rtfMRI-NF for reducing cigarette cravings and increasing smoking resistance, a number of studies have explored potential therapeutic applications of rtfMRI-NF in disorders such as Parkinson's disease (Subramanian et al., 2011), chronic pain (deCharms et al., 2005), major depression (Linden et al., 2012), schizophrenia (Ruiz et al., 2014), attention deficit hyperactivity disorder (Arns et al., 2014), and stroke (Sitaram et al., 2012). On the basis of our findings, we believe that FC-added NF would be valuable for facilitating modulation of neuronal activity and FC, particularly for neuropsychiatric disorders such as schizophrenia and attention deficit hyperactivity disorder (Suskauer et al., 2008; Hoffman et al., 1995), and thus for enhancing the associated aberrant mental functions, and this deserves extensive future investigations. Note, however, that regional activity-based NF may be well suited to certain neurological or mental disorders that are characterized by changes to localized brain regions, for example, the substantia nigra in Parkinson's disease (Aarsland, Påhlhagen, Ballard, Ehrt, & Svenningsson, 2011; Moustafa & Gluck, 2011), the subgenual cingulate cortex in major depressive disorder (Hamani et al., 2011; Cohen et al., 2009), the rostral ACC in chronic pain (Bushnell, Čeko, & Low, 2013; deCharms et al., 2005), and the region affected by focal stroke (Dirnberger, Novak, & Nasel, 2013; Sitaram et al., 2012; Särkämö et al., 2010).

Conclusion

In this study, we have presented evidence supporting the feasibility of altering neuronal networks using an rtfMRI-NF method that includes information on FC, by comparing the traditional neuronal activity-based approach to derive NF signals with a newly developed FC-added approach. The method and findings highlight the potential of the FC-added rtfMRI-NF method to allow volitional control over brain activity and connectivity patterns and to enable the modification of associated mental functions.

Acknowledgments

The authors would like to thank Mr. Han-Gil Lee, Yong-Hwan Kim, Junghoe Kim, Minkyung Oh, Eunkyung Jung, Hojung Kang, and Hyun-Chul Kim for their logistical support and earlier work in a pilot study as well as Dr. Dave Vago and Ms. Stephanie Lee for their valuable comments and edits. Preliminary results of this study were introduced as abstract at the annual workshop of the Brain Engineering Society of Korea, Jeongsun, Korea, 2013 and as conference proceeding at the International Conference on Neural Information Processing, Daegu, Korea, 2013. This work was supported by the grants from the Global Research Network program funded by NSF of Korea (NRF-2013S1A2A2035364) and from NRF/MSIP of Korea (2015R1A2A2A03004462). This work was also supported in part by a grant of the Korean Health Technology R&D Project, Ministry of Health & Welfare, Korea (HI12C1847) and in part by the BK21 Plus program funded by the NRF of Korea. M. T. is funded by the Swiss National Science Foundation (project number PZ00P1_137023). G. M. receives funding from the Swiss National Science Foundation under project no. 100014_135328. These sponsors had no involvement in study processes, including the study design, data collection, analysis/interpretation of data, writing of the report, and the decision to submit the manuscript for publication.

Reprint requests should be sent to corresponding author Jong-Hwan Lee, Department of Brain and Cognitive Engineering, Korea University, Anam-dong 5ga, Seongbuk-gu, Seoul 136-713, Republic of Korea, or via e-mail: jonghwan_lee@korea.ac.kr.

REFERENCES

Aarsland
,
D.
,
Påhlhagen
,
S.
,
Ballard
,
C. G.
,
Ehrt
,
U.
, &
Svenningsson
,
P.
(
2011
).
Depression in Parkinson disease—Epidemiology, mechanisms and management
.
Nature Reviews Neurology
,
8
,
35
47
.
Allen
,
E. A.
,
Damaraju
,
E.
,
Plis
,
S. M.
,
Erhardt
,
E. B.
,
Eichele
,
T.
, &
Calhoun
,
V. D.
(
2014
).
Tracking whole-brain connectivity dynamics in the resting state
.
Cerebral Cortex
,
24
,
663
676
.
Arns
,
M.
,
Heinrich
,
H.
, &
Strehl
,
U.
(
2014
).
Evaluation of neurofeedback in ADHD: The long and winding road
.
Biological Psychology
,
95
,
108
115
.
Ashburner
,
J.
(
2007
).
A fast diffeomorphic image registration algorithm
.
Neuroimage
,
38
,
95
113
.
Azizian
,
A.
,
Nestor
,
L. J.
,
Payer
,
D.
,
Monterosso
,
J. R.
,
Brody
,
A. L.
, &
London
,
E. D.
(
2009
).
Smoking reduces conflict-related anterior cingulate activity in abstinent cigarette smokers performing a Stroop task
.
Neuropsychopharmacology
,
35
,
775
782
.
Baumgartner
,
T.
,
Knoch
,
D.
,
Hotz
,
P.
,
Eisenegger
,
C.
, &
Fehr
,
E.
(
2011
).
Dorsolateral and ventromedial prefrontal cortex orchestrate normative choice
.
Nature Neuroscience
,
14
,
1468
1474
.
Beauchamp
,
M. S.
,
Pasalar
,
S.
, &
Ro
,
T.
(
2010
).
Neural substrates of reliability-weighted visual-tactile multisensory integration
.
Frontiers in Systems Neuroscience
,
4
,
25
.
Behzadi
,
Y.
,
Restom
,
K.
,
Liau
,
J.
, &
Liu
,
T. T.
(
2007
).
A component based noise correction method (CompCor) for BOLD and perfusion based fMRI
.
Neuroimage
,
37
,
90
101
.
Birn
,
R. M.
,
Diamond
,
J. B.
,
Smith
,
M. A.
, &
Bandettini
,
P. A.
(
2006
).
Separating respiratory-variation-related fluctuations from neuronal-activity-related fluctuations in fMRI
.
Neuroimage
,
31
,
1536
1548
.
Brodmann
,
K.
(
1909
).
Vergleichende Lokalisationslehre der Grosshirnrinde
.
Leipzig, Germany
:
Barth
.
Brody
,
A. L.
,
Mandelkern
,
M. A.
,
Olmstead
,
R. E.
,
Jou
,
J.
,
Tiongson
,
E.
,
Allen
,
V.
, et al
(
2007
).
Neural substrates of resisting craving during cigarette cue exposure
.
Biological Psychiatry
,
62
,
642
651
.
Bushnell
,
M. C.
,
Čeko
,
M.
, &
Low
,
L. A.
(
2013
).
Cognitive and emotional control of pain and its disruption in chronic pain
.
Nature Reviews Neuroscience
,
14
,
502
511
.
Canterberry
,
M.
,
Hanlon
,
C. A.
,
Hartwell
,
K. J.
,
Li
,
X.
,
Owens
,
M.
,
LeMatty
,
T.
, et al
(
2013
).
Sustained reduction of nicotine craving with real-time neurofeedback: Exploring the role of severity of dependence
.
Nicotine & Tobacco Research
,
15
,
2120
2124
.
Caria
,
A.
,
Sitaram
,
R.
, &
Birbaumer
,
N.
(
2012
).
Real-time fMRI: A tool for local brain regulation
.
The Neuroscientist
,
18
,
487
501
.
Carter
,
R. M.
,
Bowling
,
D. L.
,
Reeck
,
C.
, &
Huettel
,
S. A.
(
2012
).
A distinct role of the temporal-parietal junction in predicting socially guided decisions
.
Science
,
337
,
109
111
.
Cauda
,
F.
,
Cavanna
,
A. E.
,
D'agata
,
F.
,
Sacco
,
K.
,
Duca
,
S.
, &
Geminiani
,
G. C.
(
2011
).
Functional connectivity and coactivation of the nucleus accumbens: A combined functional connectivity and structure-based meta-analysis
.
Journal of Cognitive Neuroscience
,
23
,
2864
2877
.
Cerf
,
M.
,
Thiruvengadam
,
N.
,
Mormann
,
F.
,
Kraskov
,
A.
,
Quiroga
,
R. Q.
,
Koch
,
C.
, et al
(
2010
).
On-line, voluntary control of human temporal lobe neurons
.
Nature
,
467
,
1104
1108
.
Chai
,
X. J.
,
Castañón
,
A. N.
,
Öngür
,
D.
, &
Whitfield-Gabrieli
,
S.
(
2012
).
Anticorrelations in resting state networks without global signal regression
.
Neuroimage
,
59
,
1420
1428
.
Chai
,
X. J.
,
Ofen
,
N.
,
Gabrieli
,
J. D.
, &
Whitfield-Gabrieli
,
S.
(
2014
).
Selective development of anticorrelated networks in the intrinsic functional organization of the human brain
.
Journal of Cognitive Neuroscience
,
26
,
501
513
.
Cohen
,
M. X.
,
Axmacher
,
N.
,
Lenartz
,
D.
,
Elger
,
C. E.
,
Sturm
,
V.
, &
Schlaepfer
,
T. E.
(
2009
).
Good vibrations: Cross-frequency coupling in the human nucleus accumbens during reward processing
.
Journal of Cognitive Neuroscience
,
21
,
875
889
.
Colvin
,
M. K.
,
Dunbar
,
K.
, &
Grafman
,
J.
(
2001
).
The effects of frontal lobe lesions on goal achievement in the water jug task
.
Journal of Cognitive Neuroscience
,
13
,
1129
1147
.
Cornelius
,
J. R.
,
Lynch
,
K.
,
Martin
,
C. S.
,
Cornelius
,
M. D.
, &
Clark
,
D. B.
(
2001
).
Clinical correlates of heavy tobacco use among adolescents
.
Addictive Behaviors
,
26
,
273
277
.
Czisch
,
M.
,
Wehrle
,
R.
,
Harsay
,
H. A.
,
Wetter
,
T. C.
,
Holsboer
,
F.
,
Sämann
,
P. G.
, et al
(
2012
).
On the need of objective vigilance monitoring: Effects of sleep loss on target detection and task-negative activity using combined EEG/fMRI
.
Frontiers in Neurology
,
3
,
67
.
deCharms
,
R. C.
,
Maeda
,
F.
,
Glover
,
G. H.
,
Ludlow
,
D.
,
Pauly
,
J. M.
,
Soneji
,
D.
, et al
(
2005
).
Control over brain activation and pain learned by using real-time functional MRI
.
Proceedings of the National Academy of Sciences, U.S.A.
,
102
,
18626
18631
.
Dinn
,
W. M.
,
Aycicegi
,
A.
, &
Harris
,
C. L.
(
2004
).
Cigarette smoking in a student sample: Neurocognitive and clinical correlates
.
Addictive Behaviors
,
29
,
107
126
.
Dirnberger
,
G.
,
Novak
,
J.
, &
Nasel
,
C.
(
2013
).
Perceptual sequence learning is more severely impaired than motor sequence learning in patients with chronic cerebellar stroke
.
Journal of Cognitive Neuroscience
,
25
,
2207
2215
.
Due
,
D. L.
,
Huettel
,
S. A.
,
Hall
,
W. G.
, &
Rubin
,
D. C.
(
2002
).
Activation in mesolimbic and visuospatial neural circuits elicited by smoking cues: Evidence from functional magnetic resonance imaging
.
American Journal of Psychiatry
,
159
,
954
960
.
Ernst
,
M.
,
Matochik
,
J. A.
,
Heishman
,
S. J.
,
Van Horn
,
J. D.
,
Jons
,
P. H.
,
Henningfield
,
J. E.
, et al
(
2001
).
Effect of nicotine on brain activation during performance of a working memory task
.
Proceedings of the National Academy of Sciences, U.S.A.
,
98
,
4728
4733
.
Fox
,
M. D.
, &
Raichle
,
M. E.
(
2007
).
Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging
.
Nature Reviews Neuroscience
,
8
,
700
711
.
Fox
,
M. D.
,
Snyder
,
A. Z.
,
Vincent
,
J. L.
,
Corbetta
,
M.
,
Van Essen
,
D. C.
, &
Raichle
,
M. E.
(
2005
).
The human brain is intrinsically organized into dynamic, anticorrelated functional networks
.
Proceedings of the National Academy of Sciences, U.S.A.
,
102
,
9673
9678
.
Fox
,
M. D.
,
Zhang
,
D.
,
Snyder
,
A. Z.
, &
Raichle
,
M. E.
(
2009
).
The global signal and observed anticorrelated resting state brain networks
.
Journal of Neurophysiology
,
101
,
3270
3283
.
Fransson
,
P.
, &
Marrelec
,
G.
(
2008
).
The precuneus/posterior cingulate cortex plays a pivotal role in the default mode network: Evidence from a partial correlation network analysis
.
Neuroimage
,
42
,
1178
1184
.
Gaebler
,
M.
,
Biessmann
,
F.
,
Lamke
,
J.-P.
,
Müller
,
K.-R.
,
Walter
,
H.
, &
Hetzer
,
S.
(
2014
).
Stereoscopic depth increases intersubject correlations of brain networks
.
Neuroimage
,
100
,
427
434
.
Garrison
,
K. A.
,
Scheinost
,
D.
,
Worhunsky
,
P. D.
,
Elwafi
,
H. M.
,
Thornhill
,
T. A.
, IV
,
Thompson
,
E.
, et al
(
2013
).
Real-time fMRI links subjective experience with brain activity during focused attention
.
Neuroimage
,
81
,
110
118
.
Goebel
,
R.
,
Zilverstand
,
A.
, &
Sorger
,
B.
(
2010
).
Real-time fMRI-based brain-computer interfacing for neurofeedback therapy and compensation of lost motor functions
.
Imaging in Medicine
,
2
,
407
415
.
Haller
,
S.
,
Kopel
,
R.
,
Jhooti
,
P.
,
Haas
,
T.
,
Scharnowski
,
F.
,
Lovblad
,
K.-O.
, et al
(
2013
).
Dynamic reconfiguration of human brain functional networks through neurofeedback
.
Neuroimage
,
81
,
243
252
.
Hamani
,
C.
,
Mayberg
,
H.
,
Stone
,
S.
,
Laxton
,
A.
,
Haber
,
S.
, &
Lozano
,
A. M.
(
2011
).
The subcallosal cingulate gyrus in the context of major depression
.
Biological Psychiatry
,
69
,
301
308
.
Hamilton
,
J. P.
,
Glover
,
G. H.
,
Hsu
,
J. J.
,
Johnson
,
R. F.
, &
Gotlib
,
I. H.
(
2011
).
Modulation of subgenual anterior cingulate cortex activity with real-time neurofeedback
.
Human Brain Mapping
,
32
,
22
31
.
Hampton
,
A. N.
,
Bossaerts
,
P.
, &
O'Doherty
,
J. P.
(
2008
).
Neural correlates of mentalizing-related computations during strategic interactions in humans
.
Proceedings of the National Academy of Sciences, U.S.A.
,
105
,
6741
6746
.
Hanlon
,
C. A.
,
Hartwell
,
K. J.
,
Canterberry
,
M.
,
Li
,
X.
,
Owens
,
M.
,
LeMatty
,
T.
, et al
(
2013
).
Reduction of cue-induced craving through realtime neurofeedback in nicotine users: The role of region of interest selection and multiple visits
.
Psychiatry Research: Neuroimaging
,
213
,
79
81
.
Hanson
,
S. J.
,
Gagliardi
,
A.
, &
Hanson
,
C.
(
2009
).
Solving the brain synchrony eigenvalue problem: Conservation of temporal dynamics (fMRI) over subjects doing the same task
.
Journal of Computational Neuroscience
,
27
,
103
114
.
Hartwell
,
K. J.
,
Johnson
,
K. A.
,
Li
,
X.
,
Myrick
,
H.
,
LeMatty
,
T.
,
George
,
M. S.
, et al
(
2011
).
Neural correlates of craving and resisting craving for tobacco in nicotine dependent smokers
.
Addiction Biology
,
16
,
654
666
.
Heatherton
,
T. F.
,
Kozlowski
,
L. T.
,
Frecker
,
R. C.
, &
Fagerstorm
,
K.-O.
(
1991
).
The Fagerström test for nicotine dependence: A revision of the Fagerstrom Tolerance Questionnaire
.
British Journal of Addiction
,
86
,
1119
1127
.
Heffner
,
J. L.
,
Anthenelli
,
R. M.
,
Adler
,
C. M.
,
Strakowski
,
S. M.
,
Beavers
,
J.
, &
DelBello
,
M. P.
(
2013
).
Prevalence and correlates of heavy smoking and nicotine dependence in adolescents with bipolar and cannabis use disorders
.
Psychiatry Research
,
210
,
857
862
.
Hoffman
,
R. E.
,
Rapaport
,
J.
,
Ameli
,
R.
,
McGlashan
,
T. H.
,
Harcherik
,
D.
, &
Servan-Schreiber
,
D.
(
1995
).
A neural network simulation of hallucinated “voices” and associated speech perception impairments in schizophrenic patients
.
Journal of Cognitive Neuroscience
,
7
,
479
496
.
Holsen
,
L. M.
,
Lee
,
J.-H.
,
Spaeth
,
S. B.
,
Ogden
,
L. A.
,
Klibanski
,
A.
,
Whitfield-Gabrieli
,
S.
, et al
(
2012
).
Brain hypoactivation, autonomic nervous system dysregulation, and gonadal hormones in depression: A preliminary study
.
Neuroscience Letters
,
514
,
57
61
.
Huettel
,
S. A.
,
Song
,
A. W.
, &
McCarthy
,
G.
(
2004
).
Functional magnetic resonance imaging
(
Vol. 1
).
Sunderland, MA
:
Sinauer Associates
.
Hutchison
,
R. M.
,
Womelsdorf
,
T.
,
Allen
,
E. A.
,
Bandettini
,
P. A.
,
Calhoun
,
V. D.
,
Corbetta
,
M.
, et al
(
2013
).
Dynamic functional connectivity: Promise, issues, and interpretations
.
Neuroimage
,
80
,
360
378
.
Hutchison
,
R. M.
,
Womelsdorf
,
T.
,
Gati
,
J. S.
,
Everling
,
S.
, &
Menon
,
R. S.
(
2013
).
Resting-state networks show dynamic functional connectivity in awake humans and anesthetized macaques
.
Human Brain Mapping
,
34
,
2154
2177
.
Izuma
,
K.
,
Saito
,
D. N.
, &
Sadato
,
N.
(
2010
).
Processing of the incentive for social approval in the ventral striatum during charitable donation
.
Journal of Cognitive Neuroscience
,
22
,
621
631
.
Jackson
,
K. M.
,
Sher
,
K. J.
,
Cooper
,
M. L.
, &
Wood
,
P. K.
(
2002
).
Adolescent alcohol and tobacco use: Onset, persistence and trajectories of use across two samples
.
Addiction
,
97
,
517
531
.
Jobson
,
J. D.
(
1991
).
Applied multivariate data analysis, volume I: Regression and experimental design
.
New York, NY
:
Springer
.
Johnson
,
K. A.
,
Hartwell
,
K.
,
LeMatty
,
T.
,
Borckardt
,
J.
,
Morgan
,
P. S.
,
Govindarajan
,
K.
, et al
(
2012
).
Intermittent “real-time” fMRI feedback is superior to continuous presentation for a motor imagery task: A pilot study
.
Journal of Neuroimaging
,
22
,
58
66
.
Kim
,
D.-Y.
, &
Lee
,
J.-H.
(
2011
).
Are posterior default-mode networks more robust than anterior default-mode networks? Evidence from resting-state fMRI data analysis
.
Neuroscience Letters
,
498
,
57
62
.
Kim
,
H.-C.
,
Yoo
,
S.-S.
, &
Lee
,
J.-H.
(
2015
).
Recursive approach of EEG-segment-based principal component analysis substantially reduces cryogenic pump artifacts in simultaneous EEG-fMRI data
.
Neuroimage
,
104
,
437
451
.
Kim
,
J.
,
Kim
,
Y.-H.
, &
Lee
,
J.-H.
(
2013
).
Hippocampus–precuneus functional connectivity as an early sign of Alzheimer's disease: A preliminary study using structural and functional magnetic resonance imaging data
.
Brain Research
,
1495
,
18
29
.
Kim
,
Y. H.
,
Kim
,
J.
, &
Lee
,
J. H.
(
2012
).
Iterative approach of dual regression with a sparse prior enhances the performance of independent component analysis for group functional magnetic resonance imaging (fMRI) data
.
Neuroimage
,
63
,
1864
1889
.
Klein
,
A.
,
Andersson
,
J.
,
Ardekani
,
B. A.
,
Ashburner
,
J.
,
Avants
,
B.
,
Chiang
,
M. C.
, et al
(
2009
).
Evaluation of 14 nonlinear deformation algorithms applied to human brain MRI registration
.
Neuroimage
,
46
,
786
802
.
Kober
,
H.
,
Mende-Siedlecki
,
P.
,
Kross
,
E. F.
,
Weber
,
J.
,
Mischel
,
W.
,
Hart
,
C. L.
, et al
(
2010
).
Prefrontal–striatal pathway underlies cognitive regulation of craving
.
Proceedings of the National Academy of Sciences, U.S.A.
,
107
,
14811
14816
.
Koush
,
Y.
,
Rosa
,
M. J.
,
Robineau
,
F.
,
Heinen
,
K.
,
Rieger
,
W. S.
,
Weiskopf
,
N.
, et al
(
2013
).
Connectivity-based neurofeedback: Dynamic causal modeling for real-time fMRI
.
Neuroimage
,
81
,
422
430
.
LaConte
,
S. M.
(
2011
).
Decoding fMRI brain states in real-time
.
Neuroimage
,
56
,
440
454
.
Lee
,
J.-H.
,
Kim
,
D.-Y.
, &
Kim
,
J.
(
2012
).
Mesocorticolimbic hyperactivity of deprived smokers and brain imaging
.
NeuroReport
,
23
,
1039
1043
.
Lee
,
J.-H.
,
Kim
,
J.
, &
Yoo
,
S.-S.
(
2012
).
Real-time fMRI-based neurofeedback reinforces causality of attention networks
.
Neuroscience Research
,
72
,
347
354
.
Lee
,
J.-H.
,
O'Leary
,
H. M.
,
Park
,
H.
,
Jolesz
,
F. A.
, &
Yoo
,
S. S.
(
2008
).
Atlas-based multichannel monitoring of functional MRI signals in real-time: Automated approach
.
Human Brain Mapping
,
29
,
157
166
.
Lee
,
J.-H.
,
Ryu
,
J.
,
Jolesz
,
F. A.
,
Cho
,
Z.-H.
, &
Yoo
,
S.-S.
(
2009
).
Brain–machine interface via real-time fMRI: Preliminary study on thought-controlled robotic arm
.
Neuroscience Letters
,
450
,
1
6
.
Leonardi
,
N.
,
Richiardi
,
J.
,
Gschwind
,
M.
,
Simioni
,
S.
,
Annoni
,
J.-M.
,
Schluep
,
M.
, et al
(
2013
).
Principal components of functional connectivity: A new approach to study dynamic brain connectivity during rest
.
Neuroimage
,
83
,
937
950
.
Lewis
,
J. W.
,
Talkington
,
W. J.
,
Puce
,
A.
,
Engel
,
L. R.
, &
Frum
,
C.
(
2011
).
Cortical networks representing object categories and high-level attributes of familiar real-world action sounds
.
Journal of Cognitive Neuroscience
,
23
,
2079
2101
.
Li
,
X.
,
Hartwell
,
K. J.
,
Borckardt
,
J.
,
Prisciandaro
,
J. J.
,
Saladin
,
M. E.
,
Morgan
,
P. S.
, et al
(
2013
).
Volitional reduction of anterior cingulate cortex activity produces decreased cue craving in smoking cessation: A preliminary real-time fMRI study
.
Addiction Biology
,
18
,
739
748
.
Lin
,
F.-H.
,
McIntosh
,
A. R.
,
Agnew
,
J. A.
,
Eden
,
G. F.
,
Zeffiro
,
T. A.
, &
Belliveau
,
J. W.
(
2003
).
Multivariate analysis of neuronal interactions in the generalized partial least squares framework: Simulations and empirical studies
.
Neuroimage
,
20
,
625
642
.
Linden
,
D. E.
,
Habes
,
I.
,
Johnston
,
S. J.
,
Linden
,
S.
,
Tatineni
,
R.
,
Subramanian
,
L.
, et al
(
2012
).
Real-time self-regulation of emotion networks in patients with depression
.
PLoS One
,
7
,
e38115
.
Lindquist
,
M. A.
,
Xu
,
Y.
,
Nebel
,
M. B.
, &
Caffo
,
B. S.
(
2014
).
Evaluating dynamic bivariate correlations in resting-state fMRI: A comparison study and a new approach
.
Neuroimage
,
101
,
531
546
.
Liu
,
X.
, &
Duyn
,
J. H.
(
2013
).
Time-varying functional network information extracted from brief instances of spontaneous brain activity
.
Proceedings of the National Academy of Sciences, U.S.A.
,
110
,
4392
4397
.
Livingston
,
M. A.
,
Swan
,
J. E.
, II
,
Gabbard
,
J. L.
,
Höllerer
,
T. H.
,
Hix
,
D.
,
Julier
,
S. J.
, et al
(
2003
).
Resolving multiple occluded layers in augmented reality
.
Paper presented at the Proceedings of the 2nd IEEE/ACM International Symposium on Mixed and Augmented Reality
.
Lund
,
T. E.
,
Madsen
,
K. H.
,
Sidaros
,
K.
,
Luo
,
W.-L.
, &
Nichols
,
T. E.
(
2006
).
Non-white noise in fMRI: Does modelling have an impact?
Neuroimage
,
29
,
54
66
.
Margulies
,
D. S.
,
Kelly
,
A. M.
,
Uddin
,
L. Q.
,
Biswal
,
B. B.
,
Castellanos
,
F. X.
, &
Milham
,
M. P.
(
2007
).
Mapping the functional connectivity of anterior cingulate cortex
.
Neuroimage
,
37
,
579
588
.
Maysov
,
A.
, &
Kipyatkov
,
V. E.
(
2011
).
Contrasting dynamics of cold resistance traits in field-fresh Myrmica ants during the active season
.
Journal of Thermal Biology
,
36
,
64
70
.
Moustafa
,
A. A.
, &
Gluck
,
M. A.
(
2011
).
A neurocomputational model of dopamine and prefrontal–striatal interactions during multicue category learning by Parkinson patients
.
Journal of Cognitive Neuroscience
,
23
,
151
167
.
Mundfrom
,
D. J.
,
Perrett
,
J. J.
,
Schaffer
,
J.
,
Piccone
,
A.
, &
Roozeboom
,
M.
(
2006
).
Bonferroni adjustments in tests for regression coefficients
.
Multiple Linear Regression Viewpoints
,
32
,
1
6
.
Naqvi
,
N. H.
,
Rudrauf
,
D.
,
Damasio
,
H.
, &
Bechara
,
A.
(
2007
).
Damage to the insula disrupts addiction to cigarette smoking
.
Science
,
315
,
531
534
.
Ojima
,
S.
,
Nakamura
,
N.
,
Matsuba-Kurita
,
H.
,
Hoshino
,
T.
, &
Hagiwara
,
H.
(
2011
).
Neural correlates of foreign-language learning in childhood: A 3-year longitudinal ERP study
.
Journal of Cognitive Neuroscience
,
23
,
183
199
.
Oldfield
,
R. C.
(
1971
).
The assessment and analysis of handedness: The Edinburgh inventory
.
Neuropsychologia
,
9
,
97
113
.
Pessaux
,
P.
,
Diana
,
M.
,
Soler
,
L.
,
Piardi
,
T.
,
Mutter
,
D.
, &
Marescaux
,
J.
(
2014
).
Robotic duodenopancreatectomy assisted with augmented reality and real-time fluorescence guidance
.
Surgical Endoscopy
,
28
,
2493
2498
.
Posse
,
S.
,
Fitzgerald
,
D.
,
Gao
,
K.
,
Habel
,
U.
,
Rosenberg
,
D.
,
Moore
,
G. J.
, et al
(
2003
).
Real-time fMRI of temporolimbic regions detects amygdala activation during single-trial self-induced sadness
.
Neuroimage
,
18
,
760
768
.
Raichle
,
M. E.
, &
Snyder
,
A. Z.
(
2007
).
A default mode of brain function: A brief history of an evolving idea
.
Neuroimage
,
37
,
1083
1090
.
Rance
,
M.
,
Ruttorf
,
M.
,
Nees
,
F.
,
Schad
,
L. R.
, &
Flor
,
H.
(
2014
).
Real time fMRI feedback of the anterior cingulate and posterior insular cortex in the processing of pain
.
Human Brain Mapping
,
35
,
5784
5798
.
Rota
,
G.
,
Handjaras
,
G.
,
Sitaram
,
R.
,
Birbaumer
,
N.
, &
Dogil
,
G.
(
2011
).
Reorganization of functional and effective connectivity during real-time fMRI-BCI modulation of prosody processing
.
Brain and Language
,
117
,
123
132
.
Ruiz
,
S.
,
Buyukturkoglu
,
K.
,
Rana
,
M.
,
Birbaumer
,
N.
, &
Sitaram
,
R.
(
2014
).
Real-time fMRI brain computer interfaces: Self-regulation of single brain regions to networks
.
Biological Psychology
,
95
,
4
20
.
Ruiz
,
S.
,
Rana
,
M.
,
Sass
,
K.
,
Kircher
,
T.
,
Birbaumer
,
N.
, &
Sitaram
,
R.
(
2011
).
Brain network connectivity and behaviour enhancement: A fMRI-BCI study
.
Paper presented at the 17th Annual Meeting of the Organization for Human Brain Mapping
.
Sargent
,
J. D.
,
Morgenstern
,
M.
,
Isensee
,
B.
, &
Hanewinkel
,
R.
(
2009
).
Movie smoking and urge to smoke among adult smokers
.
Nicotine & Tobacco Research
,
11
,
1042
1046
.
Särkämö
,
T.
,
Pihko
,
E.
,
Laitinen
,
S.
,
Forsblom
,
A.
,
Soinila
,
S.
,
Mikkonen
,
M.
, et al
(
2010
).
Music and speech listening enhance the recovery of early sensory processing after stroke
.
Journal of Cognitive Neuroscience
,
22
,
2716
2727
.
Schiff
,
N. D.
,
Giacino
,
J.
,
Kalmar
,
K.
,
Victor
,
J.
,
Baker
,
K.
,
Gerber
,
M.
, et al
(
2007
).
Behavioural improvements with thalamic stimulation after severe traumatic brain injury
.
Nature
,
448
,
600
603
.
Selby
,
P.
,
Andriash
,
K.
,
Zawertailo
,
L.
,
Persad
,
D.
,
Zack
,
M.
, &
Busto
,
U. E.
(
2013
).
Escalating doses of transdermal nicotine in heavy smokers: Effects on smoking behavior and craving
.
Journal of Clinical Psychopharmacology
,
33
,
667
674
.
Shmueli
,
D.
,
Prochaska
,
J. J.
, &
Glantz
,
S. A.
(
2010
).
Effect of smoking scenes in films on immediate smoking: A randomized controlled study
.
American Journal of Preventive Medicine
,
38
,
351
358
.
Sitaram
,
R.
,
Lee
,
S.
,
Ruiz
,
S.
,
Rana
,
M.
,
Veit
,
R.
, &
Birbaumer
,
N.
(
2011
).
Real-time support vector classification and feedback of multiple emotional brain states
.
Neuroimage
,
56
,
753
765
.
Sitaram
,
R.
,
Veit
,
R.
,
Stevens
,
B.
,
Caria
,
A.
,
Gerloff
,
C.
,
Birbaumer
,
N.
, et al
(
2012
).
Acquired control of ventral premotor cortex activity by feedback training an exploratory real-time fMRI and TMS study
.
Neurorehabilitation and Neural Repair
,
26
,
256
265
.
Smith
,
A. M.
,
Lewis
,
B. K.
,
Ruttimann
,
U. E.
,
Ye
,
F. Q.
,
Sinnwell
,
T. M.
,
Yang
,
Y.
, et al
(
1999
).
Investigation of low frequency drift in fMRI signal
.
Neuroimage
,
9
,
526
533
.
Smolka
,
M. N.
,
Bühler
,
M.
,
Klein
,
S.
,
Zimmermann
,
U.
,
Mann
,
K.
,
Heinz
,
A.
, et al
(
2006
).
Severity of nicotine dependence modulates cue-induced brain activity in regions involved in motor preparation and imagery
.
Psychopharmacology
,
184
,
577
588
.
Soldati
,
N.
,
Calhoun
,
V. D.
,
Bruzzone
,
L.
, &
Jovicich
,
J.
(
2013
).
The use of a priori information in ICA-based techniques for real-time fMRI: An evaluation of static/dynamic and spatial/temporal characteristics
.
Frontiers in Human Neuroscience
,
7
,
64
.
Stein
,
E. A.
,
Pankiewicz
,
J.
,
Harsch
,
H. H.
,
Cho
,
J.-K.
,
Fuller
,
S. A.
,
Hoffmann
,
R. G.
, et al
(
1998
).
Nicotine-induced limbic cortical activation in the human brain: A functional MRI study
.
American Journal of Psychiatry
,
155
,
1009
1015
.
Stoeckel
,
L. E.
,
Garrison
,
K. A.
,
Ghosh
,
S. S.
,
Wighton
,
P.
,
Hanlon
,
C. A.
,
Gilman
,
J. M.
, et al
(
2014
).
Optimizing real time fMRI neurofeedback for therapeutic discovery and development
.
Neuroimage: Clinical
,
5
,
245
255
.
Subramanian
,
L.
,
Hindle
,
J. V.
,
Johnston
,
S.
,
Roberts
,
M. V.
,
Husain
,
M.
,
Goebel
,
R.
, et al
(
2011
).
Real-time functional magnetic resonance imaging neurofeedback for treatment of Parkinson's disease
.
The Journal of Neuroscience
,
31
,
16309
16317
.
Sulzer
,
J.
,
Haller
,
S.
,
Scharnowski
,
F.
,
Weiskopf
,
N.
,
Birbaumer
,
N.
,
Blefari
,
M. L.
, et al
(
2013
).
Real-time fMRI neurofeedback: Progress and challenges
.
Neuroimage
,
76
,
386
399
.
Suskauer
,
S. J.
,
Simmonds
,
D. J.
,
Fotedar
,
S.
,
Blankner
,
J. G.
,
Pekar
,
J. J.
,
Denckla
,
M. B.
, et al
(
2008
).
Functional magnetic resonance imaging evidence for abnormalities in response selection in attention deficit hyperactivity disorder: Differences in activation associated with response inhibition but not habitual motor response
.
Journal of Cognitive Neuroscience
,
20
,
478
493
.
Sutherland
,
M. T.
,
McHugh
,
M. J.
,
Pariyadath
,
V.
, &
Stein
,
E. A.
(
2012
).
Resting state functional connectivity in addiction: Lessons learned and a road ahead
.
Neuroimage
,
62
,
2281
2295
.
Taioli
,
E.
, &
Wynder
,
E. L.
(
1991
).
Effect of the age at which smoking begins on frequency of smoking in adulthood
.
The New England Journal of Medicine
,
325
,
968
969
.
Torta
,
D.
, &
Cauda
,
F.
(
2011
).
Different functions in the cingulate cortex, a meta-analytic connectivity modeling study
.
Neuroimage
,
56
,
2157
2172
.
Tzourio-Mazoyer
,
N.
,
Landeau
,
B.
,
Papathanassiou
,
D.
,
Crivello
,
F.
,
Etard
,
O.
,
Delcroix
,
N.
, et al
(
2002
).
Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain
.
Neuroimage
,
15
,
273
289
.
Vallesi
,
A.
,
McIntosh
,
A. R.
,
Shallice
,
T.
, &
Stuss
,
D. T.
(
2009
).
When time shapes behavior: fMRI evidence of brain correlates of temporal monitoring
.
Journal of Cognitive Neuroscience
,
21
,
1116
1126
.
Van De Ville
,
D.
,
Jhooti
,
P.
,
Haas
,
T.
,
Kopel
,
R.
,
Lovblad
,
K.-O.
,
Scheffler
,
K.
, et al
(
2012
).
Recovery of the default mode network after demanding neurofeedback training occurs in spatio-temporally segregated subnetworks
.
Neuroimage
,
63
,
1775
1781
.
Veilleux
,
J. C.
,
Conrad
,
M.
, &
Kassel
,
J. D.
(
2013
).
Cue-induced cigarette craving and mixed emotions: A role for positive affect in the craving process
.
Addictive Behaviors
,
38
,
1881
1889
.
Veit
,
R.
,
Singh
,
V.
,
Sitaram
,
R.
,
Caria
,
A.
,
Rauss
,
K.
, &
Birbaumer
,
N.
(
2012
).
Using real-time fMRI to learn voluntary regulation of the anterior insula in the presence of threat-related stimuli
.
Social Cognitive and Affective Neuroscience
,
7
,
623
634
.
Wager
,
T. D.
,
van Ast
,
V. A.
,
Hughes
,
B. L.
,
Davidson
,
M. L.
,
Lindquist
,
M. A.
, &
Ochsner
,
K. N.
(
2009
).
Brain mediators of cardiovascular responses to social threat, part II: Prefrontal-subcortical pathways and relationship with anxiety
.
Neuroimage
,
47
,
836
851
.
Wagner
,
D. D.
,
Dal Cin
,
S.
,
Sargent
,
J. D.
,
Kelley
,
W. M.
, &
Heatherton
,
T. F.
(
2011
).
Spontaneous action representation in smokers when watching movie characters smoke
.
The Journal of Neuroscience
,
31
,
894
898
.
Weiskopf
,
N.
(
2012
).
Real-time fMRI and its application to neurofeedback
.
Neuroimage
,
62
,
682
692
.
Weiskopf
,
N.
,
Scharnowski
,
F.
,
Veit
,
R.
,
Goebel
,
R.
,
Birbaumer
,
N.
, &
Mathiak
,
K.
(
2004
).
Self-regulation of local brain activity using real-time functional magnetic resonance imaging (fMRI)
.
Journal of Physiology-Paris
,
98
,
357
373
.
Weiskopf
,
N.
,
Sitaram
,
R.
,
Josephs
,
O.
,
Veit
,
R.
,
Scharnowski
,
F.
,
Goebel
,
R.
, et al
(
2007
).
Real-time functional magnetic resonance imaging: Methods and applications
.
Magnetic Resonance Imaging
,
25
,
989
1003
.
West
,
R.
,
Hajek
,
P.
,
Stead
,
L.
, &
Stapleton
,
J.
(
2005
).
Outcome criteria in smoking cessation trials: Proposal for a common standard
.
Addiction
,
100
,
299
303
.
Wilkinson
,
G. S.
, &
Robertson
,
G. J.
(
2006
).
WRAT 4: Wide Range Achievement Test; professional manual
.
Lutz, FL
:
Psychological Assessment Resources, Incorporated
.
Yoo
,
S.-S.
, &
Jolesz
,
F. A.
(
2002
).
Functional MRI for neurofeedback: Feasibility study on a hand motor task
.
NeuroReport
,
13
,
1377
1381
.
Yoo
,
S.-S.
,
Lee
,
J.-H.
,
O'Leary
,
H.
,
Lee
,
V.
,
Choo
,
S.-E.
, &
Jolesz
,
F. A.
(
2007
).
Functional magnetic resonance imaging-mediated learning of increased activity in auditory areas
.
NeuroReport
,
18
,
1915
1920
.
Yoo
,
S.-S.
,
Lee
,
J.-H.
,
O'Leary
,
H.
,
Panych
,
L. P.
, &
Jolesz
,
F. A.
(
2008
).
Neurofeedback fMRI-mediated learning and consolidation of regional brain activation during motor imagery
.
International Journal of Imaging Systems and Technology
,
18
,
69
78
.
Zalesky
,
A.
,
Fornito
,
A.
,
Cocchi
,
L.
,
Gollo
,
L. L.
, &
Breakspear
,
M.
(
2014
).
Time-resolved resting-state brain networks
.
Proceedings of the National Academy of Sciences, U.S.A.
,
111
,
10341
10346
.
Zar
,
J. H.
(
1996
).
Multiple regression and correlation
. In
J. H.
Zar
(ed.),
Biostatistical analysis
(3rd ed., pp.
353
360
).
Upper Saddle River, NJ
:
Prentice Hall
.