Being able to adjust alertness is important in daily life. We are all aware that task performance can be improved by increasing our level of alertness. Alertness appears to be influenced by various factors. A lack of sleep or tiredness can for instance lead to a lower level of alertness, which makes cognitive performance more challenging. Apparently, alertness is a function of the brain that can be modulated. Another important external factor that may influence cognitive brain activity is the demand for action. There is, however, almost no literature available from imaging research that examines this question. The goal of the current study is therefore to examine how the response frequency affects cognitive brain activity. This knowledge may aid in understanding the role of the demand for overt action on brain activity.
Arguably one of the most important discoveries that has been made using neuroimaging research is that performance of cognitive tasks does not only increase activity in specific brain regions, but also reduces activity in a large set of other brain regions [1,2]. The network of regions that reduce their activity during performance of many cognitive tasks is called the default mode network (DMN). In this study, we will focus on the effect of target frequency on both activated as well as deactivated regions of the brain.
Target detection depends on many processes other than alertness, such as visual perception, decoding, target detection, response selection and response execution. Our assumption is that the target frequency will affect the overall level of alertness during a task, and thus can be measured during presentation of targets as well as non-targets. In order to focus as much as possible on activity associated with alertness, and exclude activity related to successful target detection and motor responses, we focused on activation associated with the presentation of non-targets which did not require a response. In our experiment, we used three conditions with an increasing frequency in targets and tested for effects of target frequency in two networks, with one consisting of regions showing task-induced activation (TIA), and one of regions showing task-induced deactivation (TID).
Eighteen healthy right-handed volunteers participated in this 3T functional MRI (fMRI) study [M/F: 5/13; mean age: 28.7 years (SEM: 1.7)]. Participants were recruited through online advertisement. Exclusion criteria included: left-handedness, diagnosed neurological or psychiatric disorders, severe concussion of the brain with loss of consciousness in the past, or contraindications for the MRI-scan (such as pregnancy, magnetic elements in the body and claustrophobia). All participants gave written informed consent before the scan session. This study was approved by the Independent Ethical Committee (protocol number: NL51147.028.14).
We used a short alertness task, based on a continuous performance task (CPT) paradigm. The CPT is widely used in neuroscience research and fMRI experiments to investigate sustained attention [3–5]. In our version, participants were instructed to pay attention to a fast, sequential display of consonants. Participants were instructed only to respond if the consonant was the target letter ‘X’. In order to examine whether brain activity will be influenced by response frequency, three task conditions with an increasing number of targets were presented: low (LO) with four targets, medium (MD) with 18 targets and high (HI) with 39 targets per condition. Targets were pseudo-randomly presented such that each stimulus category had varying intervals. Each condition was presented in two 30-second blocks with 39 trials (i.e. target and non-target stimuli) within each block. The total duration of one trial was 750 ms in which the stimulus was presented for 400 ms. The task blocks were separated by 15-second rest conditions, except for the middle rest condition, which was 30 seconds. The participants were not informed about the variation in target frequency but were only instructed to respond to the target letter ‘X’.
A video projector displayed the visual stimuli on a screen placed behind the scanner, which the participants saw through a mirror attached to the head coil. The participants responded to a target by pushing a button on a button box with their right hand.
The response accuracy and reaction time were used as performance measurements. We calculated the mean accuracy and reaction time for each condition separately. To test for performance differences between the three conditions, we used two separate repeated-measures analysis of variance (ANOVA) with condition as within-subject factor for response accuracy and reaction time, respectively.
Scans were performed on a 3T Philips Achieva scanner (Philips Medical Systems, Best, the Netherlands) using a 32-channel SENSE head coil, which enables parallel imaging. A 3D T1-weighted structural image was acquired for anatomical registration purposes [scan parameters: TR/TE: 8.4/3.8 ms, FOV: 254 × 254 × 158 mm, flip angle: 8°, voxel size 1 mm isotropic, 158 slices (sagittal orientation)]. fMRI images were obtained using a 3D PRESTO pulse sequence [6,7] [scan parameters: scan time 1.5 seconds, TR/TE: 19/27 ms, FOV: 256 × 256 × 160 mm, flip angle: 10°, voxel size 4 mm isotropic, 40 slices (sagittal orientation), 201 volumes].
Functional MRI preprocessing
Functional MRI data were preprocessed and analyzed using SPM12. All scans were registered to the first scan to correct for subject movement during the experiment. Subsequently, the images were co-registered to the anatomical image and spatially normalized into standard Montreal Neurological Institute (MNI) space, using parameters derived from the spatial normalization of the anatomical image. Individual scans were spatially smoothed with a 3D Gaussian filter (full-width at half-maximum: 8 mm) to minimize effects of functional anatomical differences between subjects.
Functional MRI analysis
For the evaluation of the fMRI data, we performed an event-related general linear model (GLM) regression analysis with separate regressors for targets and non-targets in each condition, estimating the signal change compared to rest. To examine the effect of target frequency on brain activity without confounding effects of successful target detection or motor responses, we focused on results associated with the activation during non-targets. The tolerance values (which indicate the independence of regressors) were 0.53, 0.25 and 0.32 for the non-target regressor in the LO, MD and HI condition, respectively. A blocked GLM regression analysis that combined targets and non-targets for each condition was also performed in order to specify task-related activity which was used to select the regions of interest (ROIs).
Region of interest selection
A ROI analysis was performed for the group evaluation of signal changes, due to its superior statistical power over a whole-brain voxel analysis [8,9], as well as the possibility to statistically compare signal changes in different networks and regions, and because it allows for quantitative replication of the findings of our study .
A second level voxel-wise whole-brain analysis of the HI condition with targets and non-targets combined was used to determine the ROIs. Areas in which the brain activity exceeded the threshold value of |t| ≥ 3.09 were used to select ROIs. For the positive activity, we focused on activity within the central executive network [11,12], and excluded areas from our analyses that are known to be primarily related to motor-responses, such as the primary motor and premotor cortex, and excluded primary visual regions and non-cortical regions such as the thalamus. Cubic ROIs with a dimension of 15 × 15 × 15 mm were placed over the activity of these activated or deactivated brain areas (Fig. 1a). Two networks were formed with the selected ROIs, namely: TIA and TID. The average brain activity was calculated for each network. The location of the selected ROIs is presented in Fig. 1. MNI coordinates of the center point of the ROIs and the corresponding automated anatomical labeling (AAL) atlas labels  and Brodmann areas  are indicated in Table 1.
All statistical hypothesis tests were performed using SPSS 24 using the results based on activation during non-targets. We applied a repeated-measures ANOVA with the three task conditions as repeated measures for TIA and TID separately. We used the intercept test (which tests if the average signal of the conditions differs from zero) to identify if TIA or TID showed significant signal changes during non-targets. In order to test the hypothesis that the TIA or TID networks were affected by response frequency, we tested for a linear effect of target frequency in TIA and TID using a repeated measures linear contrast ANOVA. For follow-up analysis, the intercept as well as the linear contrast test were repeated for each individual ROI. Bonferroni corrections were not applied if the omnibus test of the network to which the ROI belonged was already significant.
The accuracy (mean ± SEM) was 99.8 ± 0.1, 99.6 ± 0.1 and 98.6 ± 0.4 for LO, MD and HI respectively. Reaction times (mean ± SEM) were 407 ± 10, 385 ± 9 and 400 ± 11 ms for LO, MD and HI respectively. There was no significant difference between conditions in accuracy (F = 3.28, P = 0.06) nor reaction times (F = 0.52, P = 0.48).
Region of interest analysis
Detailed results from the analyses are presented in Fig. 1 and 2 and Table 1. The TID network showed significant deactivation during non-targets compared to rest, as indicated by the significant repeated measures intercept test (F = 35.10, P < 0.001). The level of deactivation in the TID network was not affected by the target frequency, as indicated by a strongly non-significant linear contrast ANOVA test (F = 0.23; P = 0.62) (Fig. 2). Follow-up tests for the individual ROIs of the TID network confirmed these results, as all individual ROIs of the TID showed significant deactivation except for LSFGmo, LPCG and LMFG, and none of the ROIs showed a significant linear contrast (Table 1).
The TIA network did not show a significant signal change during non-targets compared to rest, as reflected by the repeated measures intercept test (F = 2.72, P = 0.12). There was also no significant signal change in the TIA network as a function of target frequency, as reflected by the linear contrast (F = 0.28, P = 0.61) (Fig. 2). Follow-up analyses on individual ROIs indicated that within the TIA network, none of the regions showed a significant signal change after correction for multiple comparisons. Furthermore, none of the individual regions showed a significant linear change with the target frequency (Table 1).
The main goal of this study was to examine if response frequency can affect cognitive brain activity. For this purpose, we designed a task based on a continuous performance paradigm, with three conditions with varying target frequency. In order to minimize the direct motor effects of the varying response frequency, we analyzed activation during non-targets that did not require a response.
Our results indicated significant changes in brain activation during non-targets in all three conditions, indicating sufficient power to potentially detect an effect of target frequency. These signal changes included predominantly deactivation. More importantly, the level of deactivation was not affected by target frequency. Our results indicate that response frequency does not affect the level of cognitive brain activity during an alertness task.
Our results may be relevant for further understanding of how the brain interacts with environmental requirements. Although it may appear intuitive that increasing the response frequency could be an adequate method to facilitate a high level of alertness, our study did not show evidence for this concept, at least not in level of brain activity. However, even if there is not a direct effect of target frequency on the level of brain activity, it may still be possible that a high response frequency has a positive effect on task performance, as it may facilitate to direct attention to relevant stimuli and not to distracting stimuli.
Our study also indicated that, apart from higher visual regions, there was only consistent deactivation during non-responses in all three conditions, but no activated regions. This deactivation is likely associated with functions that need to be inhibited to optimize performance. This finding replicates previous findings of Jansma et al.  who also did not report any increase in activation during automatic processing of non-targets. Thus, these studies suggest that concepts like attention and alertness that need to be at a high level both during targets as well as non-targets, is not reflected in increases in activation, but possibly with decreases in activation compared to rest.
An interesting finding in our study, although unrelated to our main hypotheses, was the fact that the medial frontal DMN regions were not deactivated during non-targets, despite the fact that our task required a continuous high level of alertness and goal-oriented behavior. Some previous studies have shown that the level of deactivation in the medial frontal regions is correlated with the difficulty of a task [15–18]. As the task that was used in our study was relatively simple, our results are in line with these previous findings.
Some limitations have to be taken into account while interpreting the results of this study. First, our task did not require maintenance or manipulation of information, and thus had a low level of cognitive processing. Possibly, cognitive brain activity may be affected by the response frequency in a task that requires a high level of cognitive processing. Second, while our results suggest that the deactivation plays an important role in alertness, we were not able to examine the possible involvement in performance, due to the near-perfect performance in all conditions.
To conclude, our study showed predominantly deactivation during non-responses in all three task conditions. The target frequency did not have an effect on brain activity in activated or deactivated regions during non-responses in our alertness task. This indicates that response frequency does not influence brain activity during an alertness task. Our results provide additional information relevant for the understanding of the neurophysiological implementation of cognitive control or alertness.
We would like to thank radiographers William Pigmans and Maikel Brands for their contribution to this study. This study is funded by ZonMw, a Dutch national organization for Health Research and Development (# 842003004).
Conflicts of interest
There are no conflicts of interest.
1. Shulman GL, Corbetta M, Fiez JA, Buckner RL, Miezin FM, Raichle ME, Petersen SE. Searching for activations that generalize over tasks. Hum Brain Mapp. 1997; 5:317–322
2. Raichle ME, MacLeod AM, Snyder AZ, Powers WJ, Gusnard DA, Shulman GL. A default mode of brain function. Proc Natl Acad Sci U S A. 2001; 98:676–682
3. Adler CM, Sax KW, Holland SK, Schmithorst V, Rosenberg L, Strakowski SM. Changes in neuronal activation with increasing attention demand in healthy volunteers: an fMRI study. Synapse. 2001; 42:266–272
4. Strakowski SM, Adler CM, Holland SK, Mills N, DelBello MP. A preliminary fMRI study of sustained attention in euthymic, unmedicated bipolar disorder. Neuropsychopharmacology. 2004; 29:1734–1740
5. Riccio CA, Reynolds CR, Lowe P, Moore JJ. The continuous performance test: a window on the neural substrates for attention? Arch Clin Neuropsychol. 2002; 17:235–272
6. van Gelderen P, Ramsey NF, Liu G, Duyn JH, Frank JA, Weinberger DR, Moonen CT. Three-dimensional functional magnetic resonance imaging of human brain on a clinical 1.5-T scanner. Proc Natl Acad Sci U S A. 1995; 92:6906–6910
7. Neggers SF, Hermans EJ, Ramsey NF. Enhanced sensitivity with fast three-dimensional blood-oxygen-level-dependent functional MRI
: comparison of SENSE-PRESTO and 2D-EPI at 3 T. NMR Biomed. 2008; 21:663–676
8. Poldrack RA. Region of interest analysis for fMRI. Soc Cogn Affect Neurosci. 2007; 2:67–70
9. Zandbelt BB, Gladwin TE, Raemaekers M, van Buuren M, Neggers SF, Kahn RS, et al. Within-subject variation in BOLD-fMRI signal changes across repeated measurements: quantification and implications for sample size. Neuroimage. 2008; 42:196–206
10. Kriegeskorte N, Simmons WK, Bellgowan PS, Baker CI. Circular analysis in systems neuroscience: the dangers of double dipping. Nat Neurosci. 2009; 12:535–540
11. D’Esposito M, Detre JA, Alsop DC, Shin RK, Atlas S, Grossman M. The neural basis of the central executive system of working memory. Nature. 1995; 378:279–281
12. Smith EE, Jonides J, Marshuetz C, Koeppe RA. Components of verbal working memory: evidence from neuroimaging. Proc Natl Acad Sci U S A. 1998; 95:876–882
13. Tzourio-Mazoyer N, Landeau B, Papathanassiou D, Crivello F, Etard O, Delcroix N, et al. Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage. 2002; 15:273–289
14. Brodmann K. Vergleichende Lokalisationslehre der Grobhirnrinde in ihren Prinzipien dargestellt auf Grund des Zellenbaues. 1909, Leipzig, Germany: J.A. Barth
15. Jansma JM, Ramsey NF, de Zwart JA, van Gelderen P, Duyn JH. fMRI study of effort and information processing in a working memory task. Hum Brain Mapp. 2007; 28:431–440
16. McKiernan KA, Kaufman JN, Kucera-Thompson J, Binder JR. A parametric manipulation of factors affecting task-induced deactivation in functional neuroimaging. J Cogn Neurosci. 2003; 15:394–408
17. Singh KD, Fawcett IP. Transient and linearly graded deactivation of the human default-mode network by a visual detection task. Neuroimage. 2008; 41:100–112
18. Čeko M, Gracely JL, Fitzcharles MA, Seminowicz DA, Schweinhardt P, Bushnell MC. Is a responsive default mode network required for successful working memory task performance? J Neurosci. 2015; 35:11595–11605
Keywords:© 2019 Wolters Kluwer Health | Lippincott Williams & Wilkins
alertness; continuous performance task; functional MRI; target frequency