Cognitive Systems — Differentiate attention, perception, executive functioning, language, and information processing systems

The scientific study of cognitive systems traces its roots to the intersection of philosophy, neurology, and experimental psychology. For centuries, thinkers debated whether the mind operated as a unified entity or comprised separable faculties, each responsible for distinct mental operations. The emergence of clinical neurology in the nineteenth century provided critical evidence: patients with localized brain lesions exhibited highly specific cognitive deficits—loss of speech production but preserved comprehension, impaired attention but intact memory—suggesting that cognition is not monolithic but modular. This observation catalyzed a research tradition that continues to inform contemporary neuropsychological assessment, rehabilitation, and the theoretical models tested on the EPPP.

Across this historical arc, a central question persisted: How do we carve cognition at its joints? The answer matters clinically because neuropsychological assessment, differential diagnosis, and treatment planning all require the practitioner to distinguish which cognitive system is impaired, which is preserved, and how those systems interact within the individual patient. The following sections explore each system in the depth expected for the EPPP.

Before examining each cognitive system individually, it is essential to appreciate the organizing principles that govern their study. Contemporary cognitive neuroscience holds that mental functions arise from distributed neural networks rather than single anatomical loci. A given cognitive system is best understood as a coordinated set of brain regions and pathways that are functionally specialized yet dynamically interactive. Damage or dysfunction in one system can cascade into others, producing complex clinical presentations that require nuanced assessment.

The diagram above captures a fundamental principle for the EPPP: while each cognitive system can be independently assessed and selectively impaired, no system operates in isolation. Attention serves as a prerequisite for virtually all higher-order processing; if a patient cannot sustain alertness or filter distractions, test performance across perception, language, memory, and executive measures will be globally depressed—a pattern that can mimic, but is fundamentally different from, generalized cognitive decline. Understanding these distinctions is critical for accurate differential diagnosis and for interpreting neuropsychological profiles.

Attention: Multi-Component Architecture

Posner and Petersen's influential attentional network theory (ANT) decomposes attention into three functionally and anatomically distinct networks. The alerting network maintains tonic and phasic readiness (right hemisphere, locus coeruleus norepinephrine system). The orienting network directs attention to specific locations or features (superior parietal lobule, frontal eye fields, pulvinar nucleus). The executive control network resolves conflict among competing responses (anterior cingulate cortex, lateral prefrontal cortex). Clinically, this tripartite model explains why a patient with right parietal damage may show hemispatial neglect (orienting deficit) while retaining the ability to sustain vigilance over time, or why a patient with frontal lobe damage may orient to stimuli but fail to suppress prepotent but incorrect responses.

Perception: Bottom-Up Meets Top-Down

Perception transforms raw sensory transduction into coherent mental representations through two complementary streams. Bottom-up processing (Gibson's direct perception) is data-driven: features such as edges, colors, and frequencies are detected by primary sensory cortices and assembled into objects. Top-down processing (Gregory's constructive perception) applies stored knowledge, expectations, and context to guide interpretation. The visual system exemplifies this dual architecture through its ventral ("what") and dorsal ("where/how") streams. Lesions to the ventral stream produce visual agnosia (inability to recognize objects despite intact visual acuity), while dorsal stream damage causes optic ataxia (inability to guide movements toward seen objects). Gestalt principles—proximity, similarity, closure, continuity—describe the organizational heuristics the perceptual system applies automatically.

Executive Functioning: The Conductor of Cognition

Executive functions (EF) are higher-order processes that regulate and coordinate other cognitive systems in the service of goal-directed behavior. Miyake and colleagues identified three core EF components through confirmatory factor analysis: inhibition (suppressing prepotent responses), shifting (flexibly switching between task sets), and updating (monitoring and revising working memory contents). These components are separable but moderately correlated—a principle termed unity and diversity of executive functions. The prefrontal cortex (PFC) is the principal substrate, with dorsolateral PFC subserving working memory and planning, ventromedial PFC involved in decision-making and emotional regulation, and orbitofrontal cortex mediating social cognition and reward-based learning.

Language: A Multi-Level System

Language processing is decomposed into multiple levels of analysis. Phonological processing handles the sound structure of language (superior temporal gyrus). Semantic processing extracts meaning from words and sentences (temporal-parietal junction, angular gyrus). Syntactic processing manages grammatical structure (Broca's area and left posterior temporal cortex). Pragmatic processing governs the social use of language—understanding metaphor, irony, and conversational implicature—and depends heavily on right hemisphere and frontal regions. The dual-stream model of language (Hickok & Poeppel) mirrors the visual system: a ventral stream maps sound to meaning, while a dorsal stream maps sound to articulatory motor representations.

Information Processing: The Computational Backbone

The information processing framework conceptualizes the mind as a system that encodes, transforms, stores, and retrieves information. Key parameters include processing speed (the rate at which cognitive operations are executed), capacity (the amount of information that can be held simultaneously, classically ≈7 ± 2 chunks), and automaticity (the extent to which a process requires attentional resources). Atkinson and Shiffrin's multi-store model (sensory register → short-term memory → long-term memory) provides the canonical architecture, though Baddeley's working memory model (phonological loop, visuospatial sketchpad, central executive, episodic buffer) offers a more refined account of the active maintenance and manipulation of information.

The following worked example illustrates how a clinician differentiates cognitive systems when interpreting a neuropsychological profile. This type of integrative reasoning is representative of EPPP clinical vignette questions.

Contemporary cognitive neuroscience has moved beyond the classical localizationist approach toward large-scale network models. Mesulam's model of distributed processing, for instance, proposes that complex behaviors emerge from interactions among cortical epicenters connected by white matter pathways. Similarly, Yeo and colleagues' identification of seven canonical resting-state networks via fMRI provides an empirical architecture for understanding how cognitive systems interact dynamically. The default mode network (medial PFC, posterior cingulate, angular gyrus) supports self-referential processing and is anti-correlated with the dorsal attention network during task engagement, illustrating how attention and executive systems dynamically suppress and recruit different networks depending on task demands.

For EPPP preparation, the key takeaway is that the five cognitive systems remain clinically useful constructs for organizing assessment and interpretation, even as the field moves toward more integrated models. You should be comfortable with both the classical distinctions (e.g., Broca's vs. Wernicke's aphasia) and the newer framework that understands these syndromes as resulting from disconnection within distributed language networks. Advanced topics such as connectome analysis, graph theory applications to brain networks, and dynamic causal modeling represent the cutting edge but are unlikely to appear on the exam in detail. What matters is understanding the conceptual shift from isolated modules to interactive systems.