Posner The Flanker Task Experiment – Barbara Eriksen and Charles Eriksen The

The investigation of human visual attention has long stood as one of the central pillars of cognitive psychology and cognitive neuroscience. At its core, the visual system is confronted with an ecological dilemma: the external physical environment presents an overwhelming barrage of electromagnetic radiation, high-density spatial arrays, and dynamic stimuli that vastly exceed the physiological and computational processing limits of the central nervous system. To navigate this sensory deluge without succumbing to catastrophic cognitive overload, the brain relies on selective attention. This suite of neurocognitive mechanisms allows organisms to filter irrelevant visual noise, orient sensory receptors toward behaviorally salient locations, and isolate critical target stimuli for prioritized processing and motor execution. Over the latter half of the twentieth century, the quest to isolate the temporal micro-architecture and anatomical substrates of these selective mechanisms crystallized into two foundational experimental traditions: the spatial cueing paradigm pioneered by Michael I. Posner and the continuous-flow interference paradigm introduced by Barbara A. Eriksen and Charles W. Eriksen.

The conceptual intersection of the Eriksen Flanker Task and Posner’s neuroarchitectural paradigms represents one of the most intellectually fertile syntheses in behavioral science. While Barbara and Charles Eriksen fundamentally dismantled the prevailing doctrine of early, discrete-stage perceptual filters by demonstrating that task-irrelevant peripheral distractors are automatically processed up to the level of motor response activation, Michael Posner provided the chronometric and neuroanatomical scaffolding required to parse visual selection into distinct, dissociable subsystems. Posner established that orienting to sensory events, maintaining an alert cognitive state, and resolving competition among conflicting mental representations are governed by separable yet interacting neural networks distributed across frontoparietal and subcortical pathways. When these two frameworks converged—most visibly in the modern Attention Network Test (ANT)—they furnished researchers with an unprecedented assay for interrogating the delicate balance between spatial filtering, executive control, and motor inhibition.

This comprehensive treatise examines the historical emergence, empirical mechanics, theoretical implications, neurobiological underpinnings, and computational formalisms that define the Eriksen Flanker Task and Posnerian attentional theory. By tracing how early information-processing debates transitioned from rigid structural bottlenecks to dynamic neural population vectors, continuous information flow, and parallel distributed processing, this analysis highlights the ongoing legacy of the Eriksens and Posner. Through detailed exploration of electrophysiological markers, drift-diffusion dynamics, prefrontal-cingulate-striatal circuitry, and clinical pathologies, we illuminate how the simple presentation of central targets flanked by congruous or incongruous noise transformed our understanding of the human mind under conflict.

1. Foundations of Selective Visual Attention and Early Paradigm Emergence

1.1 Historical Context of Information Processing and Visual Selection

The conceptual emergence of selective attention as a quantifiable cognitive construct was inextricably bound to the mid-twentieth-century cognitive revolution. As experimental psychology decisively broke away from the operational constraints of radical behaviorism, researchers ceased viewing human organisms merely as passive stimulus-response conduits. Instead, inspired by Claude Shannon’s mathematical formulation of information theory and Norbert Wiener’s cybernetics, the human mind began to be understood as an active, capacity-limited communications channel. Within this intellectual milieu, Donald Broadbent formulated his seminal 1958 Filter Model of attention. Broadbent posited a rigid, mechanical bottleneck situated early in the sensory processing stream. According to this architecture, raw sensory inputs undergo crude, parallel physical analysis before confronting a selective filter that permits only a single informational channel to traverse the bottleneck into the higher-capacity perceptual identification and long-term memory systems, consigning non-selected inputs to rapid, irrecoverable decay.

Broadbent’s early-selection doctrine ignited an intense, decades-long theoretical schism. Empirical vulnerabilities quickly emerged, highlighted by the “cocktail party phenomenon” and Anne Treisman’s dichotic listening experiments. Treisman demonstrated that semantically meaningful or highly salient stimuli presented on an unattended auditory channel could breakthrough into conscious awareness, leading her to formulate the Attenuation Theory. Rather than operating as an all-or-none physical gate, Treisman’s filter acted as a variable attenuator that dampened the signal-to-noise ratio of unattended information without completely obliterating its transmission. Concurrently, late-selection theorists, most prominently J. Anthony Deutsch and Diana Deutsch, advanced a radically divergent paradigm. They asserted that all sensory signals are processed automatically, pre-attentively, and in parallel up to the level of full semantic and conceptual identification. In their view, the attentional bottleneck operates late in the informational cascade, functioning not as a perceptual filter, but as an executive gateway governing access to conscious awareness, working memory storage, and motor response planning.

This theoretical standoff revealed the urgent necessity for experimental paradigms capable of isolating the temporal dynamics, spatial constraints, and representational levels at which visual selection operates. Auditory shadowing tasks, while pioneering, lacked the fine-grained spatial and chronometric precision needed to establish whether visual distractors could penetrate the cognitive architecture to prime motor programs without conscious awareness. Resolving these questions required rigorous chronometric methodologies capable of mapping the spatial distribution of visual noise and measuring the precise cost of irrelevant visual distractors on motor execution in real time.

1.2 Charles and Barbara Eriksen’s Breakthrough in Interference Dynamics

Working at the University of Illinois at Urbana-Champaign, the husband-and-wife research team of Charles W. Eriksen and Barbara A. Eriksen formulated a profound departure from traditional capacity-limited paradigms. While prior visual search methodologies required subjects to actively scan unpredictable visual arrays to locate a target among varying set sizes of distractors—thereby conflating spatial orienting latencies with perceptual identification and response selection—the Eriksens designed an elegant non-search task. Their primary objective was to bypass the confounding temporal mechanics of visual scanning and directly interrogate the structural limits of spatial selectivity, perceptual crosstalk, and stimulus-response compatibility mechanisms.

In their groundbreaking 1974 investigation, Barbara and Charles Eriksen altered the experimental landscape by presenting subjects with a target stimulus whose precise spatial coordinates were fully known prior to stimulus onset. By holding spatial uncertainty at zero, they demonstrated that even when observers are strictly instructed to direct their visual attention exclusively to a central target, physically adjacent distractor items—designated as “flankers”—could not be filtered out. The presence of flanking letters systematically modulated both response latency and accuracy as a function of the motor response mapped to those flankers.

This observation led the Eriksens to challenge the dominant conceptual framework of discrete, serial cognitive stages popularized by Franciscus Donders and Saul Sternberg. Rather than viewing cognition as a sequence of compartmentalized steps—where perceptual feature extraction must reach completion before identification begins, and identification must conclude before motor programming initiates—the Eriksens formulated what would mature into the continuous flow hypothesis. They posited that visual processing operates as an asynchronous, continuous accumulation of sensory evidence across the visual field. Sensory inputs automatically cascade into motor activation channels in real time. Consequently, peripheral noise stimuli that fall within a defined spatial proximity to the target inevitably activate their corresponding motor pathways, inducing sub-threshold response competition long before conscious perceptual resolution is attained.

1.3 Michael Posner’s Conceptualization of Visual Orienting

Simultaneously during this era of cognitive chronometry, Michael I. Posner was establishing an equally transformative empirical program at the University of Oregon. Posner operated from the premise that mental operations could be systematically decomposed into discrete elementary processes by measuring human reaction times with millisecond precision. His chronometric approach to the mind sought to construct an objective taxonomy of internal mental events, tracing the precise trajectory of information from sensory transduction to behavioral execution.

Posner’s most enduring theoretical contribution to visual attention was his formal dissociation between covert and overt visual orienting mechanisms. For decades, visual selection had been largely conflated with overt oculomotor movements—the physical reorientation of the fovea toward a stimulus via saccades. Through rigorous experimental manipulation, Posner proved that the human mind possesses an endogenous mental mechanism capable of shifting the focus of attention across the visual field entirely independently of eye movements. This covert orienting of attention functions as a mobile mental beam, dynamically pre-allocating neural processing resources to specific spatial coordinate frames prior to target onset.

To measure the operational characteristics of this covert mechanism, Posner developed the spatial cueing paradigm. By manipulating whether a spatial cue correctly indexed the future spatial location of a target (valid cue), pointed to the opposite hemifield (invalid cue), or provided no spatial localization information (neutral cue), Posner quantified the precise reaction time benefits of attentional pre-allocation and the measurable processing costs of disengaging, shifting, and re-engaging visual attention. This chronometric taxonomy laid the foundation for Posner’s early postulation of modular attentional subsystems within the human brain, establishing that visual attention is not a unitary cognitive faculty, but a collection of specialized neural networks orchestrated to support coherent interaction with the environment.

2. The 1974 Eriksen Flanker Paradigm: Design, Architecture, and Findings

2.1 Original Experimental Setup and Stimulus Configuration

The methodological architecture of Barbara and Charles Eriksen’s seminal 1974 paper, titled “Effects of noise letters upon the identification of a target letter in a nonsearch task,” was designed to eradicate spatial uncertainty and isolate interference dynamics. The experimental paradigm presented human observers with tachistoscopically displayed letter strings projected against a high-contrast background. The observer maintained central fixation on a calibrated fixation point, and the target letter was consistently presented at the exact center of the visual display. Flanking the target were task-irrelevant distractor letters arranged symmetrically along the horizontal meridian.

The critical manipulation resided in the categorical mapping of the stimulus set to discrete motor actions. The Eriksens assigned a pool of target letters (for example, the letters H and K) to an overt motor response executed by one hand (e.g., pulling a directional lever to the right, or pressing a designated right-hand telegraph key), while an alternative set of target letters (such as S and C) was mapped to an opposing motor response executed by the opposite hand (e.g., pulling a lever to the left, or pressing a left-hand key). The flanking characters were drawn from these identical stimulus categories, creating precise combinations of spatial noise surrounding the central target.

Crucially, the Eriksens parametrically varied the spatial separation between the central target and the flanking noise letters. Distances were measured in minutes of visual angle, ranging from tight spatial configurations (where flankers were separated from the target by a mere 0.08° to 0.17° of visual angle) to broader spatial distributions extending beyond 0.5° and 1.0° of visual angle. The duration of the stimulus arrays was held to brief presentations—often around 100 milliseconds—to strictly prevent the initiation of overt saccadic eye movements toward the flanking items, ensuring that all observed interference arose strictly through covert visual processing and centralized motor competition mechanisms.

2.2 Congruency Effects: Congruent, Incongruent, and Neutral Conditions

The systematic manipulation of target-flanker relationships yielded three foundational experimental conditions that define modern flanker methodology:

• Congruent (Compatible) Condition: The central target is flanked by identical letters (e.g., H H H H H) or by different letters that map to the identical motor response channel (e.g., K K H K K, where both H and K demand a right-hand response).
• Incongruent (Incompatible) Condition: The central target is flanked by distractor letters mapped to the competing, contralateral motor response channel (e.g., S S H S S, where the flanker S demands a left-hand response while the central target H requires a right-hand response).
• Neutral Condition: The central target is flanked by characters that possess no categorical mapping to either response key within the experimental context, or characters with low perceptual similarity to the target set (e.g., N N H N N or shapes, where neither stimulus has an assigned motor program).

The behavioral metrics observed in the 1974 study established the Flanker Interference Effect as one of the most robust and replicable phenomena in cognitive psychology. When observers encountered incongruent arrays, reaction times (RT) increased dramatically—often exhibiting a delay of 40 to 80 milliseconds relative to congruent trials. Furthermore, error rates rose sharply in the incongruent condition, with participants frequently executing the incorrect motor response mapped to the irrelevant flankers.

Intriguingly, the data revealed a striking asymmetry between the cost of incongruent distractors and the benefit of congruent distractors. While congruent flankers produced a modest facilitative reduction in reaction time compared to neutral baselines (often around 10 to 20 milliseconds), the processing cost imposed by incongruent flankers was disproportionately large, generating substantial RT elevation and high error rates. This asymmetry demonstrated that the motor execution system does not treat compatible and incompatible visual noise symmetrically. Incongruence introduces an acute computational conflict that demands deliberate, top-down executive intervention to abort the prematurely primed, erroneous response before the correct motor command can be executed.

2.3 Spatial Filtering, Visual Noise, and the Perceptual Window

Beyond identifying motor competition, the 1974 paradigm systematically illuminated the spatial constraints governing selective visual filtering. The Eriksens discovered that the magnitude of flanker interference was an inverse function of the spatial separation between the target and the distractors. When flankers were positioned immediately adjacent to the central target (within an envelope of approximately 0.5° of visual angle, roughly corresponding to the width of the human fovea), interference was maximal. As the lateral spacing widened beyond this critical threshold, reaction time elevations steadily decayed, eventually plateauing into a baseline where flankers exerted negligible effects on target processing.

These spatial decay functions demonstrated that the human visual system cannot instantaneously construct an infinitely sharp spatial attentional filter. Instead, covert attention operates across a continuous spatial gradient or “perceptual window.” Visual items falling within this functional window are processed in parallel, undergoing feature extraction and semantic parsing regardless of task instructions. The boundary conditions established by the Eriksens indicated that spatial selection is fundamentally constrained by spatial proximity: when distractors are situated too close to the target, the visual system fails to spatially segregate the sensory inputs early in the visual processing stream.

The Eriksens also meticulously differentiated between perceptual crosstalk—the visual confusion or lateral masking caused by physical crowding of spatial contours—and motor response competition. By utilizing neutral flankers that shared high perceptual similarity with the target letters but lacked motor response mappings, they demonstrated that while perceptual similarity does induce modest slowing via visual grouping mechanisms, it accounts for only a minor fraction of the total flanker effect. The overwhelming driver of the elevated reaction times observed in incongruent trials is the active motor competition elicited by the distractors, proving that visual selection failures propagate through the neurocognitive architecture into the premotor and motor systems.

3. Michael Posner’s Spatial Cueing Paradigm and the Components of Attention

3.1 The Tripartite Architecture of the Posner Cueing Task

While the Eriksen paradigm illuminated the spatial boundaries and motor consequences of visual noise, Michael Posner’s spatial cueing paradigm (often designated the Posner Cueing Task) provided a rigorous methodology for dissecting the mechanisms that direct the focus of attention across visual space. Posner introduced a triadic chronometric model of spatial orienting by measuring human performance under varying informational cues that preceded target onset by precise temporal intervals known as Stimulus-Onset Asynchrony (SOA).

The operational framework of the paradigm relies on the systematic manipulation of cue validity:

• Valid Cues: The cue correctly informs the participant of the exact spatial coordinates where the target will subsequently appear (typically occurring in 70–80% of trials in predictive experimental blocks).
• Invalid Cues: The cue misleadingly directs the participant’s attention to a spatial coordinate in the opposite visual field or an alternative location, where the target does not appear (typically 20–30% of trials).
• Neutral Cues: The cue provides temporal warning information regarding when the target will appear, but provides no spatial information regarding its lateral location (e.g., a central diamond, cross, or bilateral flashing borders).

Posner operationalized the difference in reaction time between valid and neutral trials as the “attentional benefit”—the processing acceleration gained from pre-allocating visual attention to the relevant spatial coordinates. Conversely, the difference between invalid and neutral trials was operationalized as the “attentional cost”—the processing penalty incurred because attention had to be actively disengaged from the incorrect location, shifted across visual space, and re-engaged at the true target locus.

Furthermore, Posner established the crucial dissociation between exogenous and endogenous orienting. Exogenous (peripheral, reflexive) orienting is triggered automatically by bottom-up, transient physical events occurring in the visual periphery, such as a sudden flash of light or abrupt luminance change. Exogenous orienting is rapid, peaking at brief SOAs between 50 and 150 milliseconds, and occurs largely outside voluntary top-down control. In contrast, endogenous (central, symbolic) orienting is driven by top-down, goal-directed processing of central, symbolic cues—such as a centrally positioned arrow or digit indicating an intended spatial target. Endogenous orienting is a voluntary, cognitively controlled operation that emerges more slowly, typically requiring 200 to 400 milliseconds of SOA to reach peak efficacy.

Crucially, at extended SOAs (typically beyond 300 milliseconds) during exogenous cueing, Posner and Yoav Cohen discovered the phenomenon known as Inhibition of Return (IOR). At these longer intervals, reaction times to previously cued, invalid peripheral locations actually become faster than to the validly cued location. IOR serves as an evolutionarily adaptive foraging mechanism, penalizing visual attention from immediately returning to previously inspected spatial coordinates and thereby encouraging the visual system to orient toward novel, uninspected regions of the physical environment.

3.2 The Three Attentional Networks: Posner and Petersen’s Formulation

Synthesizing decades of cognitive chronometry, neuroimaging, and neuropsychological lesion data, Michael Posner and Steven Petersen published a landmark 1990 framework in the Annual Review of Neuroscience. They articulated that human attention cannot be conceptualized as a unitary cognitive capacity, nor is it an emergent property of the entire cerebral cortex operating homogeneously. Instead, attention is sustained by a modular system composed of three anatomically distinct and functionally specialized neural networks:

• The Alerting Network: Responsible for achieving and maintaining an optimal state of cognitive arousal, vigilance, and sensory readiness. Phasic alerting is triggered by warning signals that prepare the cognitive apparatus for immediate impending action, while tonic alerting reflects sustained vigilance over extended time horizons. Anatomically, this network is driven by ascending noradrenergic projections originating in the pontine locus coeruleus, which innervate the right frontal and parietal cortices to adjust sensory sensitivity.
• The Orienting Network: Responsible for directing the covert and overt focus of attention to specific locations across sensory coordinate frames. This system performs three dissociable elementary mental operations: disengaging attention from a current locus, shifting the attentional focus across coordinates, and engaging the newly selected target. This network is subdivided into a dorsal frontoparietal system (including the frontal eye fields [FEF] and superior parietal lobule [SPL]/intraparietal sulcus [IPS]) governing voluntary, goal-directed orienting, and a ventral frontoparietal system (including the temporoparietal junction [TPJ] and ventral frontal cortex) acting as an attentional circuit-breaker for unexpected, salient events. Subcortical contributions involve the superior colliculus and the pulvinar nucleus of the thalamus.
• The Executive Control Network: Responsible for orchestrating complex, goal-directed behavior, detecting errors, monitoring cognitive conflict, and suppressing prepotent, automatic motor impulses in favor of deliberative action plans. The executive network resolves internal competition between rival sensory inputs and behavioral intentions. Its primary neuroanatomical hubs are anchored in the anterior cingulate cortex (ACC), the presupplementary motor area (pre-SMA), and the dorsolateral prefrontal cortex (DLPFC), with extensive reciprocal subcortical loops traversing the basal ganglia.

3.3 The Attentional Spotlight and Zoom-Lens Models

To conceptualize how spatial attention operates within physical space, Posner originally introduced the iconic “Attentional Spotlight” metaphor. In this formulation, covert visual attention mimics an internal spotlight sweeping across the visual field. Visual stimuli falling within the circumscribed beam of the spotlight receive enhanced neural processing, elevated sensory gain, and accelerated transmigration into conscious perception, whereas stimuli outside the spotlight remain unattended or severely attenuated. In Posner’s early conceptualization, the spotlight was generally considered to possess fixed spatial dimensions, moving across the visual field in an analog, continuous fashion similar to a smooth-pursuit or saccadic trajectory.

However, this rigid spotlight metaphor was fundamentally refined by Charles Eriksen and James E. St. James in their seminal 1986 paper presenting the Zoom-Lens Model. Eriksen and St. James challenged the assumption that the spatial envelope of attention maintains invariant boundaries. Instead, they demonstrated that the attentional field operates with high structural plasticity, functionally behaving like a zoom lens on a camera. Under demanding behavioral conditions requiring high spatial acuity and fine visual discrimination—such as isolating a single target letter surrounded by tight visual distractors—the attentional focus can be narrowed down to a tiny, high-density focal beam.

Critically, Eriksen and St. James established that visual attention operates under a strict, conservation-of-resources trade-off: there is an inverse relationship between the spatial extent of the attentional field and the processing efficiency (or neural resource density) per unit area. When the zoom lens is constricted to a narrow visual angle, processing capacity is densely concentrated, resulting in rapid target identification, elevated spatial resolution, and robust suppression of nearby distractors. Conversely, when the attentional focus is broadened across an expansive visual field to monitor for dispersed targets, processing efficiency drops proportionately: the signal-to-noise ratio decreases, perceptual identification slows, and peripheral flankers penetrate more deeply into cognitive processing channels.

This theoretical synthesis resolved long-standing disputes regarding early versus late selection. Spatial attention does not function as an immutable, binary spatial gate. Rather, it operates as a flexible, dynamic spatial gradient whose steepness and spatial extent are continuously modulated by task demands, expectancy, and top-down executive control.

4. The Theoretical Convergence: Integrating Posner’s Networks with the Eriksen Flanker Task

4.1 Genesis of the Attention Network Test (ANT)

For more than two decades, the empirical traditions of Posner’s spatial orienting paradigms and the Eriksen flanker interference tasks progressed largely along parallel theoretical trajectories. Posnerian researchers focused predominantly on how spatial cues shift the spatial coordinates of visual selection, while Eriksen researchers concentrated on the computational mechanics of response competition, continuous flow, and spatial noise. The definitive empirical convergence of these two paradigms occurred in 2002, when Jin Fan, Bruce McCandliss, Tobias Sommer, Amir Raz, and Michael Posner published their landmark paper introducing the Attention Network Test (ANT) in the Journal of Cognitive Neuroscience.

The ANT was engineered to provide an integrated, psychometrically sound, and chronometrically efficient experimental assay capable of measuring the operational efficiency of all three of Posner’s attentional networks—Alerting, Orienting, and Executive Control—within a single, 30-minute testing session. To achieve this synthesis, Fan and colleagues embedded the spatial and response architecture of the Eriksen Flanker Task directly within the temporal and cueing structure of Posner’s spatial cueing paradigm.

In a standard ANT trial, human observers are instructed to identify the horizontal direction of a central target arrow (pointing either left or right) by pressing a corresponding response button with their left or right index finger. On every trial, the central target is flanked bilaterally by two identical arrows on each side, forming a five-element horizontal string. The flanker arrays conform strictly to the Eriksen paradigm: they are either congruent (flankers pointing in the identical direction as the target: → → → → → or ← ← ← ← ←), incongruent (flankers pointing in the directly opposing direction: → → ← → → or ← ← → ← ←), or neutral (flankers consisting of simple horizontal lines devoid of arrowheads: — — → — —).

Crucially, the onset of these target-flanker configurations is temporally preceded by one of four distinct cue conditions derived from Posner’s cueing architecture:

• No Cue: Observers view the central fixation cross continuously without warning prior to target onset.
• Center Cue: An asterisk replaces the central fixation cross, providing temporal warning (alerting) without spatial localization information.
• Double Cue: Two asterisks appear simultaneously above and below the fixation cross, marking both possible target locations, maximizing temporal alerting while leaving spatial location uncertain.
• Spatial Cue: A single asterisk appears at the precise spatial coordinates (either above or below the fixation cross) where the target array will subsequently appear, providing both temporal alerting and definitive spatial orienting information.

By employing a fully crossed, orthogonal factorial design, the ANT allows researchers to extract the functional efficiency of each attentional network through simple arithmetic subtractions of median reaction times:

$$\text{Alerting Effect} = \text{RT}_{\text{no cue}} – \text{RT}_{\text{double cue}}$$

$$\text{Orienting Effect} = \text{RT}_{\text{center cue}} – \text{RT}_{\text{spatial cue}}$$

$$\text{Executive Control Effect} = \text{RT}_{\text{incongruent flankers}} – \text{RT}_{\text{congruent flankers}}$$

4.2 Executive Control as Flanker Conflict Resolution

The profound theoretical innovation of the ANT was the formal operationalization of Posner’s Executive Control Network specifically through the quantitative metric of Eriksen flanker interference. Prior to the ANT, executive attention was frequently probed through complex, multifaceted tasks such as the Wisconsin Card Sorting Test or the Stroop Task. However, the Stroop Task incorporates complex lexical, semantic, and phonetic transformations, whereas the Eriksen Flanker Task provides an unconfounded, purely visual-motor measure of conflict detection and prepotent response suppression.

Within the ANT framework, the Executive Control metric explicitly indexes the cognitive capacity to maintain top-down goal representation in working memory while actively suppressing bottom-up, stimulus-driven distraction. When the incongruent array appears, the flanking arrows automatically activate an incorrect motor program in the motor cortex. The executive network must detect this computational conflict, deploy inhibitory control to damp down the unauthorized motor activation vector, and bias neural gain toward the central target representation to drive the correct motor response. A larger executive subtraction score denotes lower processing efficiency in resolving conflict, reflecting an elevated cognitive cost to override lateral interference.

Moreover, the integration of these paradigms permitted researchers to observe critical, non-linear interactions across attentional networks. Psychometric and chronometric analyses revealed that the three networks, while largely anatomically and functionally dissociable, continuously modulate one another. Most notably, investigations using the ANT demonstrated that high states of alerting—elicited by warning cues—paradoxically increase the magnitude of the flanker conflict effect. When the alerting network is hyper-activated by a warning signal, the central nervous system enters a state of elevated sensory readiness and motor disinhibition. This rapid readiness accelerates the speed of initial motor channel accumulation. Consequently, the automatic, bottom-up processing of peripheral flankers is amplified, driving erroneous motor activation faster and demanding a more strenuous top-down executive effort to prevent premature response execution.

4.3 Arrow versus Letter Flankers in Cross-Paradigm Studies

While the original 1974 Eriksen paradigm utilized arbitrary letter-to-key associations (e.g., mapping arbitrary letters like H/K and S/C to discrete motor keys), the Attention Network Test and the majority of contemporary cognitive studies employ directional arrow flankers. This methodological shift introduces significant cognitive and neurofunctional distinctions that warrant rigorous theoretical differentiation.

Letter-based flanker tasks rely entirely on arbitrary, learned stimulus-response (S-R) mappings established through experimental instruction and short-term working memory maintenance. The letters themselves contain no intrinsic spatial directional meaning. Therefore, conflict in a letter flanker task arises primarily through the categorical and semantic classification of the letters, which must be decoded and mapped onto the assigned motor channels via indirect cognitive translation routes. Consequently, letter flankers often elicit longer overall reaction times and engage greater initial activity in the left-hemisphere ventral occipitotemporal processing streams dedicated to orthographic and symbolic representation.

In stark contrast, directional arrow flankers introduce intrinsic spatial compatibility effects that leverage evolutionarily ingrained, highly overlearned biological associations. Arrows possess directional spatial codes that intrinsically activate directional spatial representations. When an observer views an arrow pointing left, it automatically activates the left spatial hemifield and primes the left-hand motor cortex directly, bypassing the need for complex symbolic translation. This phenomenon closely aligns with the Simon effect, wherein task-irrelevant spatial properties of a stimulus directly trigger spatially congruent motor responses.

Consequently, arrow flanker arrays generate substantially higher degrees of automaticity and motor priming than letter variants. The conflict produced by an incongruent arrow array (e.g., → → ← → →) reflects not merely categorical competition, but an acute spatial-motor conflict where the visual array triggers conflicting hemispheric motor preparation in the primary motor cortices. The automaticity of directional decoding means that arrow flankers capture motor pathways with exceptional velocity, rendering the arrow flanker task an exceptionally potent assay for interrogating the purest temporal dynamics of prefrontal-striatal motor inhibition.

5. Cognitive and Perceptual Mechanisms Underlying Flanker Interference

5.1 Continuous Flow versus Discrete Stage Models

The foundational theoretical contribution of the Eriksen flanker paradigm to cognitive psychology was its empirical invalidation of discrete, serial stage models of human information processing. Classical cognitive models, drawing heavily on Donders’ Subtraction Method and Saul Sternberg’s Additive Factor Logic, conceptualized cognitive architecture as an assembly line of strictly sequential, encapsulated operations: Perceptual Encoding → Stimulus Identification → Response Selection → Motor Programming. Under this serial framework, the operational output of a given stage is withheld until its computational processing is complete, at which point the discrete output is transmitted in its entirety to the downstream stage.

In direct opposition to this discrete framework, Charles Eriksen and Donald Schultz formulated the Continuous Flow Hypothesis in their seminal 1979 paper. They asserted that information transmission through the cognitive architecture operates in an analog, continuous cascade. As soon as electromagnetic radiation strikes the retina and initial sensory feature extraction begins in the visual cortex, rudimentary visual information begins immediately and continuously cascading forward into downstream response selection and motor programming channels.

Under the continuous flow framework, complete stimulus identification is not a prerequisite for motor channel activation. Instead, as visual features (such as lines, curves, orientations, and edges) are incrementally accumulated, they continuously provide probabilistic evidence to competing motor execution channels. Because the visual system cannot instantaneously isolate the target from closely adjacent flankers, the sensory features of the distracting flankers are initially accumulated in parallel with those of the central target. When flankers are incongruent, they immediately begin accumulating evidence in favor of the erroneous response channel. This premature, sub-threshold motor activation occurs asynchronously and well before the perceptual identification of the central target is fully finalized, leading to detectable pre-activation of peripheral muscle groups and establishing the theoretical reality of parallel, continuous-flow mental processing.

5.2 Response Competition and Dual-Route Architectures

To mathematically and conceptually formalize how continuous sensory accumulation translates into flanker interference, cognitive scientists widely adopted Sylvan Kornblum’s Dimensional Overlap Model. Kornblum proposed that stimulus-response paradigms can be mapped across dual-route processing architectures comprising an automatic, direct processing route and an intentional, indirect goal-driven route.

When applied to the Eriksen Flanker Task, the dual-route architecture operates as follows:

• The Direct Route: Driven automatically by bottom-up sensory input. Because the flanking stimuli typically possess dimensional overlap with the assigned motor responses (both sharing directional or categorical attributes), the flankers immediately trigger the direct route. This route bypasses conscious intention, automatically pushing activation into the response channel corresponding to the flankers. This direct activation is rapid, explosive, but short-lived, decaying quickly if not sustained by conscious intent.
• The Indirect Route: Governed by top-down executive goals maintained in working memory. The indirect route specifically processes the central target, verifies task instructions, and directs activation to the correct, task-relevant motor channel. The indirect route operates with greater precision and reliability, but requires more time to achieve sufficient activation to cross the motor execution threshold.

The behavioral outcome of any flanker trial is determined by the temporal dynamics and competitive interaction of these two routes. On congruent trials, both the direct and indirect routes funnel activation into the identical motor channel, accelerating reaction times to reach the decision threshold. On incongruent trials, however, the direct route rapidly drives the incorrect response channel toward threshold. To prevent an error, the executive control system must intervene: top-down suppression must actively inhibit the erroneous direct-route activation vector, while the indirect route steadily elevates activation in the correct response channel until the motor threshold is finally reached.

This dynamic is robustly illustrated by Delta Plot Analysis. A delta plot maps the magnitude of the flanker interference effect (RT_incongruent − RT_congruent) across the entire reaction time distribution, partitioned into speed percentiles (quantiles from the fastest 10% of responses to the slowest 10%). In the Eriksen Flanker Task, delta plots uniquely demonstrate a positive slope at fast RT quantiles, followed by a pronounced leveling off or negative deflection at the slowest quantiles. The downward slope at long latencies provides direct empirical proof of late-acting executive suppression: when responses take longer to execute, top-down inhibitory control has had sufficient time to fully suppress the direct-route flanker interference, thereby reducing the net reaction time difference between incongruent and congruent trials at the tail of the distribution.

5.3 Target-Flanker Grouping and Feature Integration

Although motor response competition is the primary driver of flanker interference, the magnitude of this interference is heavily constrained and modulated by early perceptual organization and Gestalt grouping principles. The human visual cortex automatically parses incoming visual scenes by grouping elements based on classic Gestalt heuristics, most notably proximity, similarity, continuity, and common fate.

The operational impact of Gestalt grouping on flanker interference is clearly demonstrated through the Homogeneity Effect. When the flanking distractors are visually identical to one another (e.g., an incongruent array where all four flankers are identical, such as → → ← → →), they form a strong, uniform perceptual Gestalt based on similarity and common orientation. This homogenous flanker group perceptually segregates from the dissimilar central target, paradoxically forming an organized background that visually “crowds” the target and exerts potent, unified interference. Conversely, when the flanking arrays are heterogeneous (e.g., composed of mixed, dissimilar characters), the flankers fail to coalesce into a cohesive perceptual unit, weakening their collective direct-route activation and reducing the net conflict experienced by the executive system.

Furthermore, early visual segmentation mechanisms can be strategically leveraged to buffer the target from surrounding lateral noise. Introducing low-level physical visual discontinuities—such as rendering the central target in a distinct color, altering its luminance, or enclosing the target within a subtle geometric spatial frame—drastically attenuates the flanker interference effect. These physical features provide the early visual system with pre-attentive spatial segmentation cues. These cues allow the spatial attentional zoom lens to establish an immediate, high-gradient boundary around the target, effectively shielding its downstream processing stream from the intrusive sensory spillover of the flanking distractors.

6. Neurobiological Foundations: Neural Substrates of Conflict and Orienting

6.1 The Anterior Cingulate Cortex (ACC) and Conflict Monitoring

The neurobiological elucidation of how the human brain detects and processes flanker interference was revolutionized by Matthew Botvinick, Cameron Carter, Jonathan Cohen, and their colleagues through the formulation of the Conflict Monitoring Hypothesis. Integrating functional magnetic resonance imaging (fMRI) with computational modeling, this framework posits that the dorsal anterior cingulate cortex (dACC, Brodmann area 24/32) functions as an online neurocomputational alarm system that continuously monitors information flow across cognitive and motor networks for the presence of computational conflict.

Within this model, conflict is mathematically defined as the concurrent, simultaneous activation of mutually incompatible, competing neural representations or response channels. In the context of an incongruent Eriksen Flanker trial, when both the left-hand and right-hand motor execution channels are co-activated by the target and flankers respectively, energy dissipation within the motor network spikes. The dACC detects this high-energy conflict state. Extensive fMRI investigations have consistently demonstrated robust, focal blood-oxygen-level-dependent (BOLD) signal elevations within the dorsal ACC specifically during incongruent flanker trials compared to congruent and neutral baselines.

Importantly, the dACC does not itself resolve the conflict or execute the required behavioral adjustments. Rather, it serves as an evaluative monitoring hub. Upon detecting computational conflict, the dACC transmits an immediate feed-forward warning signal to the dorsolateral prefrontal cortex (DLPFC). This signal triggers the prefrontal cortex to upregulate top-down cognitive control, allocating additional attentional biasing resources toward the task-relevant target for subsequent processing cycles.

6.2 Dorsolateral Prefrontal Cortex and Top-Down Biasing

While the anterior cingulate cortex serves as the conflict detector, the dorsolateral prefrontal cortex (DLPFC, Brodmann areas 9 and 46) acts as the primary executive engine that implements top-down cognitive control. The functional architecture of the prefrontal cortex in resolving flanker interference is comprehensively articulated by Earl Miller and Jonathan Cohen’s Guided Activation Theory of prefrontal cortex function.

According to the guided activation model, the DLPFC maintains the task-specific behavioral goals, operational rules, and attentional templates in an active state within working memory (e.g., the explicit operational rule: “Respond strictly to the central arrowhead; ignore all lateral distractors”). In the presence of high flanker conflict, the DLPFC exerts top-down modulatory biasing signals that propagate down the neuroanatomical hierarchy via extensive corticocortical projections to sensory and motor cortices:

• Sensory Cortical Biasing: DLPFC efferents project to retinotopically mapped extrastriate visual areas (such as V4 and the lateral occipital complex), selectively increasing the neural firing rates of neural populations tuned to the spatial coordinates of the central target, while suppressing the receptive field gain of neurons coding for the peripheral flanker positions.
• Motor System Biasing: DLPFC outputs innervate the supplementary motor area (SMA), presupplementary motor area (pre-SMA), and primary motor cortex (M1), dampening the unauthorized motor activation vector elicited by the direct route and permitting the indirect, goal-driven motor command to achieve dominance.

The imperative role of the DLPFC in flanker conflict resolution is demonstrated by neuropsychological lesion studies. Patients suffering from focal lesions localized to the lateral prefrontal cortex exhibit catastrophic breakdowns in flanker task performance: their reaction times to incongruent arrays slow precipitously, and their error rates escalate dramatically. These individuals lose the neurobiological capacity to maintain stable top-down goal representations, rendering their motor systems acutely vulnerable to stimulus-driven capture by irrelevant visual noise.

6.3 Parietal and Subcortical Circuitry in Spatial Allocation

Visual selection during the Flanker and Posner tasks relies heavily on an intricate distributed network spanning the parietal cortices and specialized subcortical nuclei. Within the parietal lobes, the Superior Parietal Lobule (SPL) and the Intraparietal Sulcus (IPS) constitute the core hubs of the dorsal orienting network. The IPS maintains dynamic, retinotopically organized priority maps of visual space. During the presentation of an Eriksen array, the IPS coordinates the spatial tuning parameters of the attentional zoom lens, calculating the exact spatial coordinates that must be prioritized and interacting with the frontal eye fields (FEF) to hold spatial attention stationary over the central target coordinate.

Subcortically, the pulvinar nucleus of the thalamus plays a critical gating role. The pulvinar acts as a bidirectional routing switch between visual cortices and the prefrontal cortex, selectively amplifying target signals and filtering out salient visual distractors. Lesions to the pulvinar lead to severe attentional filtering deficits, wherein patients become utterly incapable of suppressing distracting flankers appearing in the visual field contralateral to the lesion.

Downstream in the motor hierarchy, the basal ganglia and specifically the subthalamic nucleus (STN) serve as a vital “motor brake” under conditions of cognitive conflict. When the anterior cingulate cortex detects the simultaneous activation of competing motor channels during an incongruent flanker trial, hyperdirect pathways project from the ACC and pre-SMA directly to the STN. The STN immediately exerts a powerful, global excitatory drive onto the internal segment of the globus pallidus (GPi), which in turn powerfully inhibits the thalamocortical motor drive. This rapid activation of the STN effectively raises the motor decision threshold—slamming on the global brakes of the motor system. This temporary pause buys the cognitive architecture crucial tens of milliseconds, preventing the immediate execution of the impulsive, direct-route error and granting the DLPFC the time required to bias the system toward the correct motor command. Deep brain stimulation (DBS) studies in patients with Parkinson’s disease targeting the STN dramatically corroborate this mechanism: high-frequency electrical disruption of the STN frequently induces severe impulsivity on incongruent flanker trials, causing patients to commit fast, uninhibited errors because the subcortical motor brake has been artificially disabled.

7. Electrophysiological Markers: ERP Dynamics in Flanker and Posner Paradigms

7.1 The N200/N2 Deflection as an Index of Perceptual and Response Conflict

Event-Related Potentials (ERPs), derived from high-density electroencephalography (EEG), provide the millisecond-level temporal resolution necessary to track the cognitive operations triggered by the Eriksen Flanker Task. Among stimulus-locked electrophysiological markers, the N200 (or N2) component stands as the definitive neural signature of conflict detection. The N2 is a negative-going potential deflection emerging over frontocentral electrode sites (maximally visible at electrodes Fz and FCz) within a temporal latency window of 200 to 350 milliseconds post-stimulus onset.

When observers are presented with incongruent flanker configurations, the amplitude of the frontocentral N2 increases substantially compared to congruent or neutral configurations. Electrophysiological dipole source localization algorithms, verified by simultaneous EEG-fMRI recordings, have traced the primary neural generator of this conflict-related N2 deflection to the midcingulate cortex and the dorsal anterior cingulate cortex (dACC). The N2 amplitude is sensitive to the amount of competition present in the visual array: as the categorical or spatial incompatibility of the flankers increases, the frontocentral N2 scales proportionately in negative amplitude.

Critically, cognitive electrophysiologists have dissociated early stimulus-level perceptual conflict from late response-level motor conflict within the N2 latency window. Through sophisticated factorial designs crossing perceptual similarity with motor compatibility, researchers demonstrated that the frontocentral N2 predominantly reflects response competition. While subtle, early posterior N2 deflections (often termed the N2pc, indexing attentional target selection over parietal-occipital sites) register initial spatial filtering dynamics, the classic frontocentral N2 serves as an electrophysiological readout of the anterior cingulate cortex registering active motor competition between rival response channels.

7.2 Error-Related Negativity (ERN/Ne) and Post-Error Positivity (Pe)

When an observer fails to successfully resolve flanker interference and inadvertently executes the incorrect response on an incongruent trial, the human brain produces one of its most remarkable electrophysiological phenomena: the Error-Related Negativity (ERN), independently discovered and described by Michael Falkenstein and colleagues, and William Gehring and colleagues in the early 1990s.

The ERN is a sharp, response-locked negative voltage deflection that peaks over frontocentral electrode sites precisely at, or within 0 to 100 milliseconds after, the overt execution of an incorrect motor response (such as an incorrect keypress). Remarkably, the ERN begins its negative trajectory prior to the completion of the physical muscle movement. This immediate timing proves that the ERN is driven by an internal, central comparator mechanism rather than by delayed peripheral sensory feedback (such as visual or proprioceptive confirmation of an error). The internal monitoring system continuously compares an efference copy of the issued motor command against the active representation of the task goal maintained in the prefrontal cortex. The moment a mismatch is detected between the executed action and the intended goal, the dorsal anterior cingulate cortex discharges the ERN.

Following the ERN, a second prominent electrophysiological component emerges: the Post-Error Positivity (Pe). The Pe is a broad, positive-going waveform peaking over centroparietal electrode sites between 200 and 500 milliseconds following an erroneous response. Unlike the ERN, which occurs automatically and can be elicited even when participants are unaware of their lapses, the Pe component is strictly dependent on conscious error awareness. The Pe reflects the conscious recognition, emotional evaluation, and subjective motivational significance of the mistake. The amplitude of the Pe component strongly correlates with the magnitude of subsequent post-error behavioral adjustments, directly predicting the degree to which an individual will exhibit compensatory cognitive slowing on the immediately subsequent trial.

7.3 Lateralized Readiness Potentials (LRP) and P300 Latency

The definitive electrophysiological vindication of Charles and Barbara Eriksen’s Continuous Flow Hypothesis over discrete stage models was delivered through the recording of the Lateralized Readiness Potential (LRP). The LRP is an electrophysiological index of unilateral motor cortex preparation, calculated by subtracting the scalp potentials over the motor cortex ipsilateral to the responding hand from the potentials recorded over the contralateral motor cortex (electrode sites C3 and C4).

In groundbreaking investigations conducted by Gabriele Gratton, Michael Coles, and Emanuel Donchin, the LRP was recorded while participants performed the Eriksen Flanker Task. On incongruent trials, the stimulus-locked LRP revealed a remarkable, diphasic waveform. Within 200 to 300 milliseconds following the presentation of an incongruent array—well before any overt behavioral response was executed—the LRP exhibited an early, “gratuitous” negative deflection toward the motor cortex corresponding to the hand mapped to the irrelevant flankers. This finding proved that the flanking distractors had triggered rapid, sub-threshold motor preparation in the incorrect hand. This incorrect motor preparation was then terminated and reversed, transitioning into a robust contralateral deflection toward the correct hand as the top-down executive system gained control.

Concurrently, analysis of the P300 (specifically P3b) component—a large positive wave peaking over parietal sites associated with the completion of stimulus evaluation and categorization—demonstrated that P300 latency remains largely invariant across congruent and incongruent flanker conditions, or exhibits only minor delays. Because the P300 latency tracks stimulus evaluation time independently of response selection, the dissociation between an invariant P300 latency and an extensively delayed behavioral reaction time confirmed that flanker interference does not significantly impair early perceptual categorization. Instead, the interference operates downstream within the motor preparation and response selection stages, just as the continuous flow architecture predicted.

8. Mathematical and Computational Modeling of Flanker Conflict

8.1 Sequential Sampling and the Drift Diffusion Model (DDM)

To transition from qualitative descriptions to formal quantitative predictions of cognitive operations, mathematical psychologists model decision-making in the Eriksen Flanker Task using sequential sampling frameworks, most notably Roger Ratcliff’s Drift Diffusion Model (DDM). The DDM conceptualizes two-choice decision processes as the continuous, stochastic accumulation of noisy sensory evidence over time, drifting between two absorbing decision boundaries until one boundary is crossed, triggering motor execution.

The standard DDM parameters encompass:

• Boundary Separation ($a$): The distance between the two decision thresholds, representing the degree of response caution or the speed-accuracy trade-off.
• Non-Decision Time ($T_{er}$): The duration consumed by early sensory transduction and late peripheral motor execution.
• Starting Point Bias ($z$): An initial bias toward one decision boundary over another.
• Drift Rate ($v$): The average rate of evidence accumulation, reflecting the quality and strength of the sensory information driving the decision.

Standard diffusion models with constant drift rates fail to adequately capture the empirical dynamics of the Eriksen Flanker Task, particularly the characteristic “fast error” tail in reaction time distributions where errors occur significantly faster than correct responses on incongruent trials. To solve this, researchers formulated time-varying drift rate extensions of the DDM. On incongruent trials, the net drift rate $v(t)$ is mathematically modeled as a dynamic function that varies across time:

$$v(t) = v_{\text{target}} – v_{\text{flankers}}(t)$$

Early in the trial ($t < 150 text{ ms}$), the flanker components dominate the drift rate because they physically outnumber the target (e.g., four flanking arrows versus one central target). This early, negative drift pushes the stochastic accumulation process downward toward the incorrect decision boundary, driving fast erroneous responses. As covert spatial attention steadily focuses onto the central target, the spatial attentional filter attenuates the flanker input ($v_{text{flankers}}(t) to 0$), and the target drift ($v_{text{target}}$) gains ascendancy, reversing the trajectory toward the correct decision boundary. This dynamic mathematical formulation accurately predicts empirical RT distributions, shape parameters, and error rates across developmental, aging, and clinical cohorts.

8.2 The Dual-Stage Two-Phase (DSTP) Model

An alternative and mathematically robust computational architecture engineered specifically to explain flanker interference is Ronald Hübner, Martin Steinhauser, and Carola Lehle’s Dual-Stage Two-Phase (DSTP) Model. The DSTP model addresses the limitations of standard sequential sampling by structurally separating early perceptual selection from response selection through a two-phase architecture:

• Phase 1: Broad Perceptual Selection. Immediately upon stimulus presentation, sensory information is accumulated in parallel across the entire visual display without spatial selectivity. Both the target and the flanking characters contribute to an initial, non-selective evidence accumulation process. On incongruent trials, this phase is dominated by the flankers, rapidly feeding evidence toward the incorrect response channel in the response selection stage.
• Phase 2: Focused Attentional Selection. Simultaneously, a secondary, independent spatial filtering process runs to locate and isolate the central target. Once this spatial filtering mechanism completes its selection, the cognitive system transitions into Phase 2. The non-selective accumulation of the distractors is terminated, and the evidence accumulation stream is driven exclusively by the central target features.

The DSTP model mathematically formalizes these interactions through discrete diffusion processes operating across both stages. By implementing this structural transition from broad parallel accumulation to focused spatial selection, the DSTP model provides exact mathematical fits for empirical delta plots. Specifically, it accurately replicates the famous late-stage negative slope of delta plots, formalizing how the completion of spatial filtering suppresses lateral interference late in the trial trajectory and mathematically capturing the dynamics of the continuous flow hypothesis.

8.3 Connectionist and Neural Network Simulations

Complementing sequential sampling models, artificial intelligence and cognitive science have deployed Connectionist and Parallel Distributed Processing (PDP) models to simulate flanker interference at the level of interacting neural population vectors. Landmark models formulated by Jonathan Cohen, Michael Servan-Schreiber, and James McClelland utilize multi-layered artificial neural networks comprising input layers, hidden feature representation layers, task-demand modules, and competitive response output layers linked via recurrent inhibitory connections.

In these connectionist architectures, units interact via continuous differential equations. The presentation of an incongruent flanker array activates competing nodes within the output layer. These rival response nodes engage in mutual lateral inhibition—an attractor network dynamic where each node attempts to suppress its competitor. The conflict between these nodes generates elevated network energy, which is modeled as an input to a specialized “conflict monitoring unit” simulating the anterior cingulate cortex.

More recently, biophysically realistic, spiking neural network models have advanced these simulations by incorporating detailed cortical-striatal-thalamic-cortical loops. In these neurocomputational models, spiking pyramidal neurons in the DLPFC provide top-down excitatory currents that maintain task goals, medium spiny neurons within the striatum process competing motor options, and subthalamic nucleus populations deliver fast hyperdirect inhibition to pause motor output. Reinforcement learning algorithms (such as actor-critic architectures driven by simulated dopamine release) govern trial-to-trial synaptic weight modifications, reproducing human-like trial-by-trial behavioral adaptations, error corrections, and the precise electrophysiological waveforms of the ERN and N2 components.

9. Sequential Effects, Trial-by-Trial Adaptations, and Contextual Modulation

9.1 The Gratton Effect and Conflict Adaptation

One of the most consequential discoveries in cognitive control research was documented by Gabriele Gratton, Michael Coles, and Emanuel Donchin in their 1992 paper. They revealed that the magnitude of the Eriksen Flanker interference effect is not static across an experimental block, but undergoes continuous, dynamic modulation based on the trial-by-trial sequence of congruency conditions. Specifically, they observed that the reaction time cost of an incongruent flanker trial is significantly smaller if the preceding trial was also incongruent ($iI$ trials), compared to when the preceding trial was congruent ($cI$ trials). Conversely, performance on a congruent trial is faster following a congruent trial ($cC$ trials) than following an incongruent trial ($iC$ trials). This sequential phenomenon is known as the Gratton Effect or Conflict Adaptation.

Under the Conflict Monitoring Hypothesis, the Gratton Effect is interpreted as an index of dynamic, online cognitive control adjustment. When an observer encounters an incongruent trial, the high computational conflict detected by the dorsal anterior cingulate cortex (dACC) triggers an immediate, transient upregulation of top-down cognitive control by the dorsolateral prefrontal cortex (DLPFC). Consequently, for the subsequent trial, the prefrontal cortex maintains heightened attentional biasing: the spatial zoom lens is pre-emptively constricted around the central target locus. If the subsequent trial is also incongruent, the pre-constricted attentional filter successfully blocks the flanking distractors from invading the processing stream, thereby minimizing flanker interference.

However, this high-level executive control account has faced intense theoretical challenge from low-level associative learning perspectives, most notably articulated by Bernhard Hommel and Ulrich Mayr. They demonstrated that the Gratton Effect is frequently confounded by feature integration and episodic stimulus-response retrieval. When an identical stimulus-response configuration repeats across successive trials (e.g., an exact repetition of the letter string S S H S S), performance benefits from low-level, pre-attentive repetition priming. Conversely, partial feature repetitions (where the target changes but the flankers repeat, or vice versa) induce binding costs that mimic executive adaptation. Rigorous modern paradigms that completely eliminate feature repetitions have demonstrated that while low-level episodic priming accounts for a sizable portion of the variance, an authentic, residual cognitive conflict adaptation effect persists, verifying that the human executive control network executes genuine, trial-by-trial adjustments of attentional gain.

9.2 Proportion Congruency and Context-Dependent Control

The flexibility of the executive control network in modulating the spatial attentional filter is further highlighted by Proportion Congruency Effects. When researchers systematically alter the statistical probability of encountering conflict within an experimental block, the cognitive apparatus adapts its baseline control state accordingly:

• List-Wide Proportion Congruency (LWPC): If an experimental block consists predominantly of incongruent trials (e.g., 80% incongruent, 20% congruent), observers maintain an enduring, proactive control strategy. The spatial attentional zoom lens remains continuously constricted around the central target coordinate, dramatically reducing the flanker interference effect throughout the entire block. Conversely, in blocks where incongruent trials are rare (e.g., 20% incongruent, 80% congruent), observers adopt a relaxed, expansive attentional filter, leaving them vulnerable to massive flanker interference when an unexpected incongruent trial occurs.
• Item-Specific Proportion Congruency (ISPC): Control can also be tuned to specific visual items. If specific flanker letters (e.g., S and C) are paired with incongruence 80% of the time, while other letters are paired with incongruence only 20% of the time, observers automatically retrieve a constricted attentional filter selectively upon the physical presentation of the high-conflict items.
• Context-Specific Proportion Congruency (CSPC): Demonstrates that top-down control states can be bound to environmental contexts, such as the spatial location of the stimulus (e.g., incongruence occurring frequently at the top of the display and rarely at the bottom) or visual background colors.

These sophisticated context-dependent adaptations underscore the architectural depth of Posner’s executive control network. The brain does not merely respond reactively when conflict strikes; it builds predictive statistical models of environmental conflict distributions, proactively adjusting its spatial filters and motor thresholds to minimize computational interference.

9.3 Affective, Motivational, and Arousal Influences

Attentional filtering and conflict resolution in the Eriksen and Posner paradigms are profoundly intertwined with affective state, motivational incentives, and autonomic nervous system arousal. Grounded in neurochemical dynamics, acute biological stress and autonomic arousal heavily modulate the functional operations of the alerting and executive networks.

Elevations in phasic noradrenergic arousal, mediated by locus coeruleus projections firing in response to acute stressors, profoundly reshape the spatial envelope of the attentional zoom lens. While moderate arousal sharpens executive control, extreme stress triggers an uninhibited, impulsive processing mode: the sensory gain of peripheral distractors is amplified, baseline motor thresholds are depressed, and flanker interference escalates. Conversely, motivational incentives—such as contingent monetary rewards offered for fast, accurate responses—elicit immediate, top-down prefrontal optimization. Functional imaging reveals that reward expectancy activates the ventral striatum and dopaminergic midbrain, which directly upregulate DLPFC firing, leading to enhanced visual cortex biasing, tighter spatial filtering, and a marked reduction in flanker interference.

Furthermore, emotional valence exerts potent bottom-up capture effects. When the neutral or directional elements of an Eriksen task are replaced with, or flanked by, emotionally evocative stimuli—such as threat-related, fearful, or angry facial expressions—the amygdala automatically triggers a rapid processing shortcut through subcortical pulvinar pathways. Threat-related flankers completely bypass voluntary top-down spatial filters, capturing the attentional apparatus and producing massive, asymmetrical interference profiles that severely degrade central target resolution.

10. Developmental Trajectories and Life-Span Alterations in Attentional Control

10.1 Ontogeny of Executive Attention in Infancy and Childhood

The functional maturation of the three attentional networks defined by Posner and interrogated via the Eriksen paradigm exhibits marked developmental heterochrony. While the basic reflexive components of the Orienting Network mature early in human ontogeny—becoming functional within the first 6 to 12 months of life—the Executive Control Network undergoes a prolonged, protracted developmental trajectory that extends through childhood, adolescence, and into early adulthood.

To measure these developmental milestones in pediatric cohorts, Jin Fan, Bruce McCandliss, and colleagues adapted the adult ANT into the Child Attention Network Test (Child ANT), colloquially known as the “Fish Flanker Task.” In this colorful adaptation, the central arrow is replaced by an animated yellow fish, and the flanking arrows are replaced by lateral fish swimming in either the identical (congruent) or opposite (incongruent) direction. Children are instructed to feed the hungry central fish by pressing a directional button corresponding to the direction its mouth is facing.

Empirical studies utilizing the Child ANT reveal that young children (ages 4 to 7) exhibit massive flanker interference effects, often manifesting reaction time costs exceeding 150 to 200 milliseconds alongside error rates exceeding 30% on incongruent trials. This vulnerability stems from the delayed structural maturation of the prefrontal cortex, which undergoes prolonged myelination, synaptic pruning, and dopaminergic receptor stabilization throughout childhood. The developmental curves demonstrate that while spatial orienting reaches near-adult efficiency by age seven, the capacity to resolve flanker conflict and suppress prepotent motor impulses refines steadily into late adolescence. Longitudinal studies confirm that individual performance on the Flanker task during early childhood serves as a powerful predictive behavioral biomarker for long-term academic achievement, mathematical reasoning capabilities, and socio-emotional self-regulation skills.

10.2 Cognitive Control in Healthy Aging and Senescence

At the opposite end of the human lifespan, the aging process imposes differential degradation across the architecture of selective attention. One of the dominant theoretical frameworks explaining age-related cognitive decline is Lynn Hasher and Rose Zacks’ Inhibitory Deficit Hypothesis, which posits that normal senescence is characterized by a selective vulnerability in the neurobiological mechanisms responsible for suppressing irrelevant sensory information and inhibiting prepotent behavioral responses.

When evaluated using the Attention Network Test and the Eriksen Flanker Task, older adults (aged 65+) demonstrate a pronounced dissociation across networks:

• Preservation of Orienting: Basic covert spatial orienting in response to Posnerian cues remains remarkably intact in healthy aging, indicating that the parietal and subcortical pathways governing spatial shifts are preserved against early neurodegenerative attrition.
• Selective Vulnerability of Executive Control: Flanker interference metrics reveal dramatic degradation. Older adults exhibit severely elevated incongruent reaction times, an amplified cost of distractor interference, and a marked reduction in post-error compensatory slowing.

Neurobiologically, these behavioral deficits map directly onto age-related frontostriatal volume loss, gray matter thinning within the dorsal anterior cingulate cortex, and progressive loss of white matter tract integrity along the superior longitudinal fasciculus and the corpus callosum. To cope with these biological constraints, the aging brain frequently exhibits functional compensatory recruitment. Functional imaging studies reveal that while younger adults resolve flanker conflict using strongly lateralized right-prefrontal networks, high-performing older adults display bilateral prefrontal cortex activation—a compensatory neural phenomenon formalized as the HAROLD (Hemispheric Asymmetry Reduction in Older Adults) model.

10.3 Cognitive Reserve and Lifestyle Interventions

Despite the structural challenges imposed by biological aging, life-span research has uncovered significant plasticity in the executive control network, driven by lifestyle interventions and experiential factors that build Cognitive Reserve.

A prominent and vigorously debated area of research concerns the Bilingual Advantage. Ellen Bialystok and colleagues demonstrated that lifelong bilingual individuals, who must continuously manage two competing linguistic systems by activating one language while actively suppressing the other, exhibit enhanced conflict resolution efficiency on the Eriksen Flanker Task. Lifelong bilinguals frequently demonstrate significantly smaller flanker interference effects and reduced error rates compared to matched monolingual peers, with these cognitive benefits persisting robustly into late senescence.

Similarly, regular physical aerobic exercise has been shown to exert direct neuroprotective and neurogenic effects on the neural substrates of conflict resolution. Longitudinal randomized controlled trials demonstrate that older adults enrolled in aerobic exercise regimens exhibit increased gray matter volume in the anterior cingulate cortex and hippocampus, accompanied by significant reductions in Eriksen flanker interference and normalized ERN amplitudes. Additionally, specialized training paradigms, ranging from action video game play (which broadens the spatial deployment of the attentional zoom lens without sacrificing central resolution) to computerized cognitive training, indicate that targeted engagement of the frontoparietal control network can drive meaningful neuroplastic adaptations throughout the human lifespan.

11. Clinical Neuropsychiatry: Paradigms as Probes of Neurocognitive Dysfunction

11.1 Attention-Deficit/Hyperactivity Disorder (ADHD)

The Eriksen Flanker Task and the Posner Attention Network Test serve as critical neurocognitive probes for interrogating the pathophysiology of Attention-Deficit/Hyperactivity Disorder (ADHD). Clinically conceptualized as a disorder of executive dysfunction and behavioral disinhibition, ADHD manifests robustly in altered flanker performance profiles.

Individuals diagnosed with ADHD consistently display an exaggerated flanker interference effect, characterized not only by elevated mean reaction times and excessive error rates on incongruent trials, but crucially by pronounced intra-individual response time variability. When reaction time distributions are fitted with ex-Gaussian mathematical models, patients with ADHD exhibit massive increases in the exponential tail parameter ($tau$), reflecting frequent, periodic “attentional lapses” where top-down prefrontal control collapses completely, allowing lateral flankers to drive impulsive responses.

Electrophysiologically, patients with ADHD demonstrate marked reductions in the amplitude of the Error-Related Negativity (ERN) and attenuated frontocentral N2 conflict waveforms. Functional neuroimaging traces these electrophysiological deficits to profound hypoactivation of the frontostriatal network, including the dorsal anterior cingulate cortex, the dorsolateral prefrontal cortex, and the caudate nucleus. Pharmacotherapy with psychostimulants, such as methylphenidate or amphetamine compounds, acts directly on these circuits by blocking dopamine and norepinephrine reuptake transporters. This elevated catecholamine availability within the synaptic clefts of the ACC and DLPFC normalizes frontostriatal BOLD signals, reinstates the structural integrity of the ERN waveform, and systematically restores flanker conflict resolution efficiency toward neurotypical benchmarks.

11.2 Schizophrenia and Psychotic Disorders

In schizophrenia, the computational breakdown of cognitive architecture is severe and pervasive. Probed through the lens of Posnerian and Eriksen paradigms, schizophrenia exhibits a distinct functional dissociation that clarifies the boundaries between sensory orienting and executive control.

When administered the Posner spatial cueing paradigm, patients with schizophrenia generally display intact automatic, exogenous spatial orienting, proving that low-level subcortical and parietal reflexive mechanisms remain functional. However, when confronted with the Eriksen Flanker Task, their performance deteriorates catastrophicly. Patients demonstrate massive flanker interference, profound behavioral slowing, and an almost complete abolition of the Gratton effect (conflict adaptation). They are unable to utilize the conflict experienced on trial $n-1$ to constrict their attentional filter for trial $n$.

This deficit reflects a profound pathology in context processing and top-down task set representation, driven by NMDA-receptor hypofunction and gamma-aminobutyric acid (GABAergic) interneuron dysfunction within the dorsolateral prefrontal cortex. This neurochemical lesion disrupts task-evoked high-frequency gamma-band (30–80 Hz) neural oscillations, which are essential for coordinating long-range synchrony between the prefrontal cortex, the anterior cingulate, and visual sensory areas. Because these functional communication loops are fractured, patients cannot maintain stable goal representations to shield the central target from distracting sensory noise. Due to its high heritability and stability across clinical states, flanker interference degradation is widely recognized as a viable endophenotypic cognitive biomarker for genetic liability to schizophrenia spectrum disorders.

11.3 Affective, Anxiety, and Obsessive-Compulsive Disorders

Psychiatric disorders characterized by dysregulated affective processing and repetitive behaviors yield distinct, divergent neurofunctional profiles on the Eriksen Flanker Task:

• Obsessive-Compulsive Disorder (OCD): OCD patients present a unique, paradoxical electrophysiological profile. Rather than showing executive deficits, individuals with OCD display hyperactive conflict monitoring. When completing the Flanker task, OCD patients exhibit pathologically exaggerated, hyper-intense ERN and frontocentral N2 amplitudes. The anterior cingulate cortex in OCD behaves like a hypersensitive alarm system that perpetually registers error and conflict even during benign processing states. This hyperactive monitoring manifests behaviorally as post-error over-compensation and extreme response caution.
• Anxiety Disorders: In generalized anxiety and social anxiety disorders, the presence of threat-related distractors or environmental uncertainty drastically magnifies flanker interference. Attentional Control Theory posits that anxiety impairs the processing efficiency of the central executive network by diverting working memory capacity toward threat monitoring, leaving insufficient resources available to suppress lateral flankers.
• Major Depressive Disorder (MDD): Depression is characterized by psychomotor slowing alongside a marked blunting of the Error-Related Negativity (ERN). Depressed individuals fail to recruit adaptive post-error compensatory slowing following incongruent flanker errors, reflecting an apathy of the internal performance monitoring system and an impaired capacity of the prefrontal-striatal axis to update behavioral policy in real time. Consequently, researchers routinely deploy the Flanker task and the ANT as objective chronometric assays to track the restoration of cognitive control following interventions such as repetitive Transcranial Magnetic Stimulation (rTMS) and electroconvulsive therapy (ECT).

11.4 Neurodegenerative Pathology: Parkinson’s and Alzheimer’s Diseases

In the domain of neurodegenerative illness, the integration of Posner’s spatial metrics with Eriksen’s conflict indices provides invaluable differential diagnostic utility, particularly in distinguishing between frontostriatal and temporoparietal pathologies.

In Parkinson’s Disease (PD), degeneration of the dopaminergic substantia nigra pars compacta disrupts the basal ganglia circuitry that regulates motor execution thresholds. When tested on the Eriksen Flanker Task, PD patients off dopaminergic medication exhibit profound difficulty in suppressing the motor interference generated by incongruent flankers, alongside severe deficits in shifting attention across Posnerian spatial coordinates. Intriguingly, when PD patients receive high doses of dopamine replacement therapy or deep brain stimulation (DBS) of the subthalamic nucleus, their motor slowing improves, but their performance on high-conflict flanker trials frequently becomes profoundly impulsive. Pharmacological or electrical over-stimulation bypasses the subthalamic motor brake, allowing direct-route flanker activation to breach the motor threshold and triggering rapid, uninhibited errors.

Conversely, in early-stage Alzheimer’s Disease (AD) and amnestic Mild Cognitive Impairment (MCI), neurofibrillary tangle deposition and amyloid pathology initiate in the entorhinal cortex and spread rapidly into the posterior parietal cortices and temporoparietal junctions. Consequently, AD patients demonstrate early, catastrophic breakdowns in the Posner Orienting Network—specifically in the mental operation of disengaging covert attention from an invalid spatial cue. In contrast, performance on basic Eriksen flanker interference tasks can remain relatively spared during the earliest stages of the disease, providing a clear chronometric dissociation that assists clinicians in differentiating Alzheimer’s dementia from Dementia with Lewy Bodies or Frontotemporal Dementia, where executive flanker conflict resolution is compromised at the earliest clinical presentation.

12. Methodological Nuances, Experimental Confounds, and Future Horizons

12.1 Methodological Artifacts and Experimental Confounds

While the Eriksen Flanker Task is one of the most widely implemented paradigms in cognitive neuroscience, rigorous empirical investigation requires vigilant management of critical methodological artifacts and experimental confounds that can distort theoretical interpretations.

Foremost among these is the pervasive confound between higher-order executive adaptation and low-level associative priming. As established in critiques of the Gratton effect, standard flanker sequences frequently introduce unconstrained stimulus-response repetitions. When identical targets or flankers repeat across consecutive trials, performance benefits do not necessarily reflect prefrontal cognitive control upregulation, but rather episodic feature integration and negative priming mechanisms. To rigorously isolate genuine executive conflict adaptation, modern researchers must implement expanded stimulus pools (such as 4-choice or 8-choice letter/arrow sets) that completely prevent the immediate repetition of physical stimuli and response keys across adjacent trials.

A second critical confound resides in low-level sensory physiology: namely, the conflation of flanker interference with visual crowding and visual acuity drops across retinal eccentricity. Visual crowding is an early, pre-attentive sensory bottleneck occurring in visual areas V1 through V4, where the physical features of adjacent items are involuntarily pooled and blended together, rendering individual characters unrecognizable. If flanking distractors are positioned too close to the target, or placed deep into the peripheral visual field where receptive fields are large, the observed reaction time costs may reflect visual crowding rather than central motor response competition. Researchers must therefore carefully calibrate stimuli using cortical magnification factors ($M$-scaling) and preserve spacing outside Bouma’s bound (the critical distance threshold of crowding, typically half the eccentricity value) to ensure that the measured interference originates from central cognitive networks rather than early sensory bottlenecks.

Finally, researchers must contend with the psychometric consequences of Speed-Accuracy Trade-Offs (SATO) and hardware-induced timing jitter. Alterations in instructional framing (e.g., instructing subjects to “respond as fast as possible” versus “respond as accurately as possible”) dynamically manipulate the diffusion model’s boundary separation parameter ($a$), which can artificially inflate or suppress the raw millisecond magnitude of the flanker effect. Furthermore, in the contemporary era of online psychological testing, variations in consumer monitor refresh latencies, display buffering, and USB input device polling rates introduce substantial milliseconds of timing noise, necessitating rigorous sub-millisecond precision hardware synchronization during electrophysiological and neuroimaging protocols.

12.2 Contemporary Neuroimaging and Neurostimulation Paradigms

The methodologies utilized to investigate the Posner and Eriksen paradigms have evolved far beyond the behavioral reaction time measurements of the 1970s. Modern cognitive neuroscience relies heavily on simultaneous EEG-fMRI recordings. By recording high-density scalp electrophysiology concurrently inside the magnetic resonance bore, researchers achieve unprecedented spatiotemporal precision. This allows them to correlate the millisecond-by-millisecond trial-level fluctuations of the frontocentral N2 and ERN amplitudes directly with localized blood-oxygen-level-dependent (BOLD) hemodynamic signals in the dorsal anterior cingulate cortex, frontoparietal networks, and basal ganglia nuclei, charting the complete spatio-temporal cascade of conflict detection and resolution.

In clinical neurosurgical populations, researchers employ high-density intracranial electrocorticography (ECoG) and stereotactic depth electrodes (sEEG) in awake human patients performing the flanker task. These direct brain recordings reveal explosive, localized bursts of high-gamma power (70–150 Hz) occurring directly within the dorsal ACC and pre-SMA exactly 200 milliseconds post-stimulus, providing ground-truth verification of local population codes driving conflict detection.

Concurrently, non-invasive neurostimulation techniques provide the causal leverage that correlational imaging lacks. Applying repetitive Transcranial Magnetic Stimulation (rTMS) or continuous theta-burst stimulation (cTBS) over the left or right dorsolateral prefrontal cortex temporarily disrupts top-down executive control, directly increasing the magnitude of the flanker interference effect. Conversely, delivering transcranial direct current stimulation (tDCS) over the anterior cingulate cortex can enhance conflict monitoring, sharpening behavioral adjustments and accelerating response times. Furthermore, cognitive pupillometry has emerged as a rich, non-invasive metric: the continuous recording of task-evoked pupil diameter—a physiological proxy for locus coeruleus-norepinephrine system activity—allows researchers to track the millisecond deployment of cognitive effort and autonomic resource recruitment as the brain resolves flanker incongruency in real time.

12.3 Artificial Intelligence, Deep Learning, and Computational Attention

In the contemporary era of artificial intelligence, the empirical architectures established by Michael Posner, Barbara Eriksen, and Charles Eriksen serve as critical benchmarks for evaluating and advancing artificial neural networks. Over the past decade, deep learning has been transformed by the formulation of Transformer Architectures governed by artificial “Attention Mechanisms.” However, these computational mechanisms differ fundamentally from biological visual attention.

When state-of-the-art Deep Convolutional Neural Networks (DCNNs) and Vision Transformers (ViTs) are evaluated on synthetic visual displays matching the Eriksen Flanker Task, standard feedforward architectures consistently fail to replicate human-like performance profiles. Artificial computer vision models frequently exhibit extreme vulnerability to peripheral visual noise or, conversely, fail entirely to exhibit human-like continuous flow dynamics, spatial zoom-lens scaling, and motor response competition. Feedforward networks lack the recurrent, feedback architectures that characterize the human visual hierarchy.

To bridge this divide, AI researchers are actively constructing bio-inspired recurrent neural network (RNN) architectures that incorporate biological principles of attention:

• Lateral Inhibition: Implementing horizontal inhibitory connections within artificial convolutional layers to mimic the local suppressive circuits of the primary visual cortex, attenuating adjacent noise.
• Top-Down Recurrent Modulatory Loops: Embedding bidirectional feedback pathways that allow higher-level goal-representation layers to iteratively modulate the sensory feature gain of low-level input layers, directly replicating the DLPFC-extrastriate biasing dynamics observed during flanker conflict resolution.
• Dual-Route Gating Networks: Designing modular networks that separate fast, direct sensory-motor mappings from slow, deliberative reasoning pathways, capturing the human-like continuous information accumulation and speed-accuracy trade-offs formalized by the continuous flow hypothesis.

These bio-inspired advancements not only elevate the robustness and noise-tolerance of machine vision systems, but also grant cognitive scientists powerful computational testbeds to simulate complex lesion patterns and explore the evolutionary utility of selective attention.

Conclusion

The profound contributions of Barbara A. Eriksen, Charles W. Eriksen, and Michael I. Posner fundamentally transformed our understanding of human cognition. By dismantling the early-twentieth-century dogma of rigid, passive sensory filters and serial processing assembly lines, their experimental paradigms opened the black box of the human mind to millisecond-level chronometric precision. Barbara and Charles Eriksen proved that selective visual attention is fundamentally constrained by spatial proximity and that task-irrelevant environmental stimuli continuously and automatically invade downstream motor channels. In doing so, they revealed the computational reality of continuous flow, response competition, and motor inhibition. In parallel, Michael Posner illuminated the modular nature of the attentional architecture, mapping the distinct behavioral trajectories and anatomical pathways of the Alerting, Orienting, and Executive Control networks.

When these intellectual paradigms converged in frameworks such as the Attention Network Test, they provided cognitive science with one of its most potent, enduring, and versatile experimental assays. Across five decades of empirical research, the synthesis of the Flanker task and Posnerian theory has illuminated the neurocomputational functions of the anterior cingulate cortex, mapped the top-down biasing dynamics of the prefrontal cortex, uncovered the electrophysiological reality of the N2 and Error-Related Negativity, and provided rigorous computational foundations for sequential sampling and drift-diffusion models. Furthermore, these paradigms continue to yield profound clinical breakthroughs, guiding the diagnosis and therapeutic tracking of neurodevelopmental, psychiatric, and neurodegenerative disorders.

As cognitive science advances into an era defined by high-density intracranial recordings, simultaneous multimodal neuroimaging, and bio-inspired artificial intelligence, the conceptual frameworks pioneered by the Eriksens and Posner remain as vital today as they were in the mid-twentieth century. By revealing how the human brain deftly balances the spatial orienting of sensory receptors with the rigorous executive suppression of lateral noise, their work stands as an enduring monument in the scientific endeavor to understand how the mind generates coherent, goal-directed behavior out of the chaotic sensory tapestry of the physical world.

References

• Botvinick, M. M., Braver, T. S., Barch, D. M., Carter, C. S., & Cohen, J. D. (2001). Conflict monitoring and cognitive control. Psychological Review, 108(3), 624–652. https://doi.org/10.1037/0033-295X.108.3.624
• Broadbent, D. E. (1958). Perception and communication. Pergamon Press. https://doi.org/10.1037/10037-000
• Deutsch, J. A., & Deutsch, D. (1963). Attention: Some theoretical considerations. Psychological Review, 70(1), 80–90. https://doi.org/10.1037/h0039515
• Eriksen, B. A., & Eriksen, C. W. (1974). Effects of noise letters upon the identification of a target letter in a nonsearch task. Perception & Psychophysics, 16(1), 143–149. https://doi.org/10.3758/BF03203267
• Eriksen, C. W., & Schultz, D. W. (1979). Information processing in visual search: A continuous flow conception and experimental results. Perception & Psychophysics, 25(4), 249–263. https://doi.org/10.3758/BF03198804
• Eriksen, C. W., & St. James, J. D. (1986). Visual attention within and around the field of focal attention: A zoom lens model. Perception & Psychophysics, 40(4), 225–240. https://doi.org/10.3758/BF03211502
• Falkenstein, M., Hohnsbein, J., Hoormann, J., & Blanke, L. (1991). Effects of crossmodal divided attention on late ERP components. II. Error processing in choice reaction tasks. Electroencephalography and Clinical Neurophysiology, 78(6), 447–455. https://doi.org/10.1016/0013-4694(91)90062-9
• Fan, J., McCandliss, B. D., Sommer, T., Raz, A., & Posner, M. I. (2002). Testing the efficiency and independence of attentional networks. Journal of Cognitive Neuroscience, 14(3), 340–347. https://doi.org/10.1162/089892902317361886
• Gehring, W. J., Goss, B., Coles, M. G., Meyer, D. E., & Donchin, E. (1993). A neural system for error detection and compensation. Psychological Science, 4(6), 385–390. https://doi.org/10.1111/j.1467-9280.1993.tb00586.x
• Gratton, G., Coles, M. G., & Donchin, E. (1992). Optimizing the use of information: Strategic control of activation of responses. Journal of Experimental Psychology: General, 121(4), 480–506. https://doi.org/10.1037/0096-3445.121.4.480
• Hasher, L., & Zacks, R. T. (1988). Working memory, comprehension, and aging: A review and a new view. Psychology of Learning and Motivation, 22, 193–225. https://doi.org/10.1016/S0079-7421(08)60041-9
• Hübner, R., Steinhauser, M., & Lehle, C. (2010). A dual-stage two-phase model of selective attention. Psychological Review, 117(3), 759–784. https://doi.org/10.1037/a0019471
• Kornblum, S., Hasbroucq, T., & Osman, A. (1990). Dimensional overlap: Cognitive basis for stimulus-response compatibility–a model and taxonomy. Psychological Review, 97(2), 253–270. https://doi.org/10.1037/0033-295X.97.2.253
• Miller, E. K., & Cohen, J. D. (2001). An integrative theory of prefrontal cortex function. Annual Review of Neuroscience, 24(1), 167–202. https://doi.org/10.1146/annurev.neuro.24.1.167
• Posner, M. I. (1980). Orienting of attention. Quarterly Journal of Experimental Psychology, 32(1), 3–25. https://doi.org/10.1080/00335558008248231
• Posner, M. I., & Petersen, S. E. (1990). The attention system of the human brain. Annual Review of Neuroscience, 13(1), 25–42. https://doi.org/10.1146/annurev.ne.13.030190.000325
• Ratcliff, R. (1978). A theory of memory retrieval. Psychological Review, 85(2), 59–108. https://doi.org/10.1037/0033-295X.85.2.59
• Treisman, A. M. (1964). Selective attention in man. British Medical Bulletin, 20(1), 12–16. https://doi.org/10.1093/oxfordjournals.bmb.a070274