Interacting sources of interference during sensorimotor integration processes

Highlights

• The interrelation of different sources of interference is examined.

• We account for interactive effects of different sources of interference.

• Theta band power was modulated by interactive inference effects.

• The underlying functional neuroanatomical network is shown.

Abstract

Every day, a multitude of interfering sensory inputs needs to be integrated and adequately processed using response selection processes. Interference effects are typically investigated using classical paradigms like the Flanker and Simon task. The sources of interference for Flanker and Simon effect are known to be different and according to dual process accounts, two distinct functional pathways are involved in resolving these types of interference. It is an open question how far these sources of interference are related to each other and interact. We investigated this question in a system neurophysiological study utilizing a hybrid paradigm combining both Flanker effect-like and Simon effect-like features. We focus on event-related theta oscillations and use beamforming methods to examine functional neuroanatomical networks. The results show that Simon and Flanker interference interacted in a non-additive fashion by modulating theta band activity, probably reflecting the recruitment of cognitive control processes. Beamforming source reconstruction revealed that theta band activity was related to a broad neuronal network comprising prefrontal and cerebellar regions, including the MFG, SFG, IFG, and SMA. These regions were connected to interference processing and conflict resolution, but differed in the amount of specificity for different sources of interference.

Introduction

Our everyday environment constantly confronts us with a multitude of sensory inputs that need to be transformed into appropriate motor responses. The required sensorimotor integration and response selection processes can be complicated when it is difficult to map sensory information to the appropriate response. In cognitive psychology, such stimulus–response compatibility effects have been subject to intense research in the past decades (e.g., Fitts and Seeger, 1953, Hommel, 1997, Ivanoff et al., 2014, Kornblum et al., 1990, Kornblum et al., 1999, Leuthold, 2011). These compatibility effects are frequently investigated using simple reaction time tasks, like the “Flanker task” or the “Simon task” (e.g., Eriksen and Eriksen, 1974, Eriksen, 1995, Keye et al., 2013, Simon, 1990). In all these paradigms, an irrelevant stimulus feature can facilitate the correct response by activating the same response as the relevant stimulus feature (non-conflicting trial), or elicit a response tendency opposite to the correct response (conflicting trial) (Keye et al., 2013). However, the source of interference leading to response conflict is different for these tasks (e.g., Keye et al., 2013, Mansfield et al., 2013). In the Simon task, people make choice responses on the basis of a non-spatial stimulus feature (e.g., a letter or form). Response keys are arranged on a horizontal spatial dimension, and the location of the non-spatial stimulus feature (e.g., letter or form) varies along the same spatial dimension. Conflict effects are higher if task-irrelevant stimulus location and response button position mismatch than if stimulus location and response button position correspond. By contrast, the Flanker task requires people to make choice responses on the basis of a spatial stimulus feature (e.g., an arrow pointing towards to the left or right), but this feature is not shifted along the horizontal dimension. Instead, one or more irrelevant stimuli that are equipped with the same spatial feature are presented alongside the task-relevant target stimulus. Therefore, the source of interference and emerging conflict during sensorimotor processes is due to different sources of information; i.e., stimulus features in the Flanker Task and spatial features in the Simon task.

In line with the notion of separate sources of interference, it has been shown that inference and the processes required to resolve it involve different neurophysiological processes for Simon and Flanker tasks (Mansfield et al., 2013). Other studies have shown that conflict and interference effects are larger for Flanker than for Simon stimuli (Stins et al., 2005, Stoffels and van der Molen, 1988). From a theoretical point of view, interference effects have been suggested to result from a dual-process (e.g., De Jong et al., 1994, Ridderinkhof et al., 1995). It has been assumed that there is direct unconditional perceptual–response activation by features shared between stimuli and responses. The second process is a conditional selection of the relevant feature(s) and the appropriate response due to the stimulus–response binding (e.g., left-pointing arrow = left button press). The differential magnitude of conflict effects in the Flanker and Simon task may be due to the importance of unconditional processes in the Simon task and due to conditional processes in the Flanker task (Mansfield et al., 2013).

Given the evidence suggesting that Flanker and Simon interference effects are distinct from each other and emerge due to distinct cognitive mechanisms, a question standing to reason is in how far these sources of interference (i.e., the conditional and the unconditional route) relate to each other. To examine how Flanker and Simon interferences relate to each other and whether there are additive or non-additive (i.e., interactive) effects of different sources of interference, we used a novel hybrid paradigm combining the features of both Flanker and Simon tasks. In case of additive effects, the Simon interference effect should be similarly modulated by compatible Flanker (cFlanker) and incompatible Flanker (iFlanker) information. However, it is also possible that there are non-additive (interactive) effects: In the case of interactive effects, it is likely that when incompatible Simon (iSimon) information coincides with iFlanker information, the conflict effect (i.e., slowing of response times, decreases in response accuracy) is larger, as compared to a combination where iSimon information coincides with iFlanker information. Hence, it is also conceivable that the conflict effect is smallest in conditions where compatible Simon (cSimon) information coincides with cFlanker information.

To examine the neuronal mechanisms underlying the interrelation of different sources of conflict on a systems level, we focus on event-related neural oscillations in the theta frequency band and use beamforming methods to investigate the underlying neuronal networks. This is done for a number of reasons: Modulations of theta frequency band power have been related to cognitive control processes (Cavanagh et al., 2012, Cavanagh and Frank, 2014, Cohen, 2014, De Blasio and Barry, 2013) that are known to be increasingly required in conflict situations (e.g., Botvinick et al., 2001). Moreover, frontal theta activity can be observed whenever an unexpected uncertainty of correct behavior occurs (i.e., a conflict), most likely signaling the need of cognitive control (Cavanagh and Frank, 2014). Theta oscillations are supposed to establish cognitive control by means of coupling spatially distal sites in the brain (Fries, 2005) to organize relevant neural processes (Buzsáki and Draguhn, 2004, Uhlhaas et al., 2010, Womelsdorf et al., 2010, Anguera et al., 2013, Cohen et al., 2009, Hanslmayr et al., 2008, Nigbur et al., 2012). Consequently, modulations of theta band activity have frequently been observed during monitoring of conflicting information (Beste et al., 2010, Beste et al., 2011, Beste et al., 2012, Cohen and Donner, 2013, Lavallee et al., 2014, Tang et al., 2013, Wang et al., 2014). We hypothesize that theta band activity is modulated in similar direction as the behavioral data obtained in the combined Flanker–Simon paradigm. Concerning the functional neuroanatomy, we expect that brain regions including the middle and medial frontal gyrus, as well as the superior frontal gyrus to be associated with modulations in theta band activity. These brain regions have frequently been shown to be involved in conflict monitoring processes (Cavanagh et al., 2012, Cavanagh and Frank, 2014, Cohen, 2014, Cohen and Donner, 2013). We hypothesize that the interaction of Flanker-related and Simon-related conflicting information occurs within areas of the middle, medial and the superior frontal gyrus. Given that Flanker interference is due to conditional processes and Simon effects are due to unconditional processes (Mansfield et al., 2013), this would suggest that before-mentioned functional neuroanatomical regions may be the regions where these routes interact.