Decreased occipital alpha oscillation in children who stutter during a visual Go/Nogo task
Introduction
In developmental stuttering speech is dysfluent due to interruptions, repetitions and prolongations that complicate communication, often causing anxiety or fear of speaking. Many theories have associated stuttering severity with temperamental factors as high emotional reactivity (Conture et al., 2006, Bloodstein and Bernstein Ratner, 2008, Eggers et al., 2010) and self-regulation and inattention (Eggers et al., 2012, Eggers et al., 2013, Eggers and Jansson-Verkasalo, 2017; please see also review by Alm, 2014). Although the biological origin of stuttering is still uncertain, recent imaging studies have shown widespread structural and functional brain abnormalities in individuals who stutter when compared to fluently speaking controls (Belyk et al., 2015, Beal et al., 2007, Brown et al., 2005, Budde et al., 2014, Chang et al., 2009, Giraud et al., 2008, Jansson-Verkasalo et al., 2014, Neef et al., 2011, Preibisch et al., 2003, Salmelin et al., 2000, Sommer et al., 2002, Sowman et al., 2017, Watkins et al., 2008; for an overview, see reviews by Alm, 2004, Neef et al., 2015, Etchell et al., 2017). In addition or instead of implicating particular cortical areas, there is evidence of reduced connectivity in various white matter tracts such as tracts between auditory and motor areas, the corpus callosum or the frontal aslant tract (Cai et al., 2014, Civier et al., 2015, Kronfeld-Duenias et al., 2016). Similar, but not identical brain irregularities have been discovered in children who stutter (CWS) (Beal et al., 2013, Chang et al., 2008, Chang and Zhu, 2013, Chang et al., 2015, Misaghi et al., 2018). Chang et al. also recently described atypical functional brain network connections involving attentional and executive control related networks in CWS in a large fMRI study (Chang et al., 2017).
Considering the abovementioned findings, the CWS are likely to exhibit also attentional and inhibitory control problems. Posner and Petersen divided the attentional systems to subsystems (Posner and Petersen, 1990) that operate through several interacting functional brain networks (Visintin et al., 2015, Xuan et al., 2016; see review by Petersen and Posner, 2012). The vigilance network maintains alertness while the orienting network selects the crucial information produced by sensory systems. The executive control system, on the other hand, provides top-down control and deals e.g. with conflicting responses. A part of executive control, inhibitory control is a key factor in the regulation of impulsivity and enables attending to relevant stimuli only and thus accurate responses (Eggers and Jansson-Verkasalo, 2017, Rothbart, 1989, Rothbart and Posner, 1985).
We recently explored the inhibitory control of children who stutter by utilizing a visual Go/Nogo paradigm with simultaneous EEG and event-related potential measurement (Piispala et al., 2016a, Piispala et al., 2016b). In this task the Go-signal is to be reacted to while the Nogo-signal requires inhibition of response. In our studies the CWS showed significantly delayed Go-N2 component peak latency (Piispala et al., 2016a) as well as atypically distributed and prolonged Nogo-N2 and diminished Nogo-P3 components when compared to typically developed children (TDC) (Piispala et al., 2016b). These findings indicate problems in stimulus evaluating and classification and response selection as well as inhibitory control, despite similar accuracy and reaction times in the task.
However, ERPs can only give a narrow representation of the phase-locked neural activity in the task since ERPs usually contain parallel activity on multiple oscillatory frequencies, such as the alpha and theta band oscillations. With time-frequency analysis it is possible to gain a fuller insight into the spectral dynamics involved in the Go/Nogo task (Cavanagh and Frank, 2014, Cooper et al., 2016, Harper et al., 2014, Kirmizi-Alsan et al., 2006). Alpha and theta are the major pre-stimulus oscillations that individually affect the main ERP amplitudes (De Blasio and Barry, 2013b, De Blasio and Barry, 2013a). High pre-stimulus alpha activity increases the P3 amplitudes in a Go/Nogo task independent of condition (De Blasio and Barry, 2013b). On the other hand, low pre-stimulus theta power produced higher Nogo-N2 and Go-P3, but reduced Nogo-P3, linking low pre-stimulus theta to improved cognitive processing (De Blasio and Barry, 2013a). Theta oscillation has been shown to specifically participate in the generation of the N2 component in Go/Nogo tasks (Cavanagh et al., 2012, Harper et al., 2014). In contrast, beta frequencies only influenced the early exogenous components (De Blasio and Barry, 2013b) and delta affected all components globally (De Blasio and Barry, 2013a).
Alpha-band oscillation between 8 and 12 Hz is the dominant rhythm in the brain (Berger, 1929, Klimesch, 1999, Klimesch, 2012). By many theories, alpha modulation operates as an attentional suppression and control mechanism operating via inhibition; high alpha activity or synchronization inhibits the processing of competing irrelevant information or distractors (Jensen and Mazaheri, 2010, Foxe and Snyder, 2011, Klimesch et al., 2007; see also reviews by Freunberger et al., 2011, Frey et al., 2015). Desynchronization of alpha oscillation is usually seen in the brain areas responsible for the processing of relevant, attended information (see reviews by Frey et al., 2015, Klimesch, 2012). Pre-stimulus alpha modulation targets attentional resources to the need-to-be-attended stimulus when cues are used. In those situations alpha is desynchronized over areas processing the attended stimuli and enhanced on other, task-irrelevant areas (Slagter et al., 2016; see reviews by Frey et al., 2015, Klimesch, 1999). In particular, high alpha amplitude or synchronization prior to a visual stimulus correlates negatively with visual perception and discrimination while desynchronization predicted good performance (Van Dijk et al., 2008; see also review by Hanslmayr et al., 2011).
Theta band is usually defined as oscillations within the 4–7 Hz frequency range and it has been connected to many cognitive functions, e.g. encoding new information, learning and working memory function (Chaieb et al., 2015; see reviews by Benchenane et al., 2011, Freunberger et al., 2011, Klimesch, 1999). In regard to inhibitory control, theta oscillation particularly in the frontal midline area may be crucial in cognitive control (Kirmizi-Alsan et al., 2006, Nigbur et al., 2011; for an overview, see review by Cavanagh and Frank, 2014).
Few studies have implemented EEG and time-frequency analysis in people with stuttering and mostly using speech paradigms (Sengupta et al., 2016, Sengupta et al., 2017, Salmelin et al., 2000, Mersov et al., 2016). In a resting state EEG study increased connectivity between motor speech and premotor areas at theta and alpha range oscillations was correlated to stuttering severity (Joos et al, 2014). Metzger et al. implemented a motor Go/Nogo task and discovered atypical network activation involving the basal ganglia in the preparation of the task in adults who stutter (Metzger et al., 2018).
These aforementioned studies used adult subjects, who may actually show major compensatory changes in response to years of dysfluency. Studies on children are thus highly valuable as they represent the early years of the clinical symptoms. Therefore, in this study on 7–9 year old children who stutter, we extended the evaluation of the EEG data collected previously and additional data during a visual Go/Nogo task (Piispala et al., 2016a, Piispala et al., 2016b) from time domain ERP-analysis to time-frequency domain analysis in order to further differentiate the underlying parallel neuronal activations. We also expanded the exploration to a wider time frame including the pre-stimulus and late post-stimulus time windows to identify possible cue-related and preparatory as well as resting state differences. Considering the association between the N2 and P3 ERP-components and alpha and theta oscillations in inhibitory tasks, it is likely that the groups differ also by the magnitude, ratio and/or distribution of these oscillations. We focused on alpha and theta oscillations for their particular and relatively well-defined role in attention and cognitive control. So far there are no studies that we are aware of, on brain oscillations related to the Go/Nogo task in CWS.
Section snippets
Participants
Twelve children with stuttering (mean age 7.97 years, range 6.3–9.5 years; right-handed boys) and 12 typically developed, fluently speaking boys (mean age 8.01 years, range 5.8–9.6 years; one left-handed) participated in the study. The stuttering group consisted of 11 children already included in our previous studies (Piispala et al., 2016a, Piispala et al., 2016b) and one new subject. In order to prevent sex-related confound only boys from the previous control group were now included.
Results
Statistical testing between groups showed CWS group to present significantly suppressed (p = .014) alpha activity across a broad time window starting from around 600 ms after the Nogo - stimuli (Fig. 2). Decreased alpha activity cluster comprised mainly of parieto-occipital and frontal electrodes (for full electrode layout result please see the Supplementary Material). In Go - condition the difference between the groups was not statistically significant.
The visual FFT analysis of the chosen
Discussion
In our study the TDC showed distinct alpha band activity over the occipital areas from around 900 ms post-stimulus when the visual stimulus had ended. However, in the CWS group this was significantly reduced and virtually no alpha was seen. Occipital alpha rhythm is an expected observation in eyes-closed resting-state EEG, but is also seen in eyes-open EEG although smaller in amplitude (Barry et al., 2007, Barry et al., 2009). When judged by visual analysis of our wavelet graphs as well as the
Caveats and confounds
One consideration is the effect of the processing of the task stimuli. As in Nogo condition, clear alpha activity was indeed evident in the TDC in the same time window in the Go condition, too. However, the difference to the CWS did not reach statistical significance. The motor response in the Go condition may affect the oscillatory activity even in this later time window obscuring the differences in the statistical analysis.
On the other hand, we cannot rule out the use of some compensating
Conclusion
Our findings of significant differences in the key components of the brain’s oscillatory activity in a non-speech related setting further stress the fact that stuttering is more than a speech related disorder. Our results express the electrophysiological correlates of the extensive brain deviations found in earlier imaging studies and support the role of EEG analysis as a useful and sensitive tool in discovering these group-level differences.
Acknowledgements
We are grateful to all the children and the families who participated in this study. We also thank M.Sc. Risto Bloigu from Medical Informatics and Statistics Research Group, the University of Oulu, technician Raija Remes, systems specialist Hannu Wäänänen and M. Sc. Kalervo Suominen from the Department of Clinical Neurophysiology, Oulu University Hospital, for their help in performing and reporting of this study.
Financial disclosure
This study was supported by the Academy of Finland (grant 128840) and general research funding from the Finnish Government, granted by Oulu University Hospital. The funding sources did not have any role in the design of the study, collection of data or the analysis, interpretation or reporting of the data.
Conflict of interest
None of the authors have potential conflicts of interest to be disclosed.
References (101)
Stuttering and the basal ganglia circuits: a critical view of possible relations
J Commun Disord
(2004)Stuttering in relation to anxiety, temperament, and personality: review and analysis with focus on causality
J Fluency Disord
(2014)- et al.
A review of electrophysiology in attention-deficit/hyperactivity disorder: I. Qualitative and quantitative electroencephalography
Clin Neurophys
(2003) - et al.
EEG differences between eyes-closed and eyes-open resting conditions
Clin Neurophys
(2007) - et al.
EEG differences in children between eyes-closed and eyes-open resting conditions
Clin Neurophys
(2009) - et al.
A voxel-based morphometry (VBM) analysis of regional grey and white matter volume abnormalities within the speech production network of children who stutter
Cortex
(2013) - et al.
Oscillations in the prefrontal cortex: a gateway to memory and attention
Curr Opin Neurobiol
(2011) - et al.
EEG development of healthy boys and girls. Results of a longitudinal study
Electroencephalogr Clin Neurophysiol
(1984) - et al.
Stuttering, induced fluency, and natural fluency: a hierarchical series of activation likelihood estimation meta-analyses
Brain Lang
(2014) - et al.
Frontal theta as a mechanism for cognitive control
Trends Cogn Sci
(2014)