Cognitive Functioning

Hartmann’s (1939/1958) concept of the conflict-free ego sphere identified a domain of psychological functioning that operates — under ordinary conditions — outside drive–defense conflict. Perception, memory, attention, motility, and language were not, on this account, primarily organized by conflict. They were adaptive apparatuses that emerge in development and serve the organism’s engagement with reality.

That claim mattered. It authorized the direct measurement of cognitive functions as functions, rather than as mere epiphenomena of drive or defense. Bellak, Hurvich, and Gediman later operationalized a related ego-functioning tradition; the present battery inherits that permission structure.

It does not inherit Hartmann’s model as the architecture of the tasks.

Hartmann named the functions. He did not specify the stores, the time constants, the capacity limits, or the control processes that move information from one stage to the next. Those specifications come from information-processing psychology. The clinical vocabulary of primary autonomy, secondary autonomy, and “reinvasion” remains useful as a description of when an ostensibly automatic process becomes affectively loaded. The measurement model of how the process runs is Atkinson–Shiffrin and Baddeley.

Atkinson and Shiffrin (1968) proposed that human memory is not a single faculty but a system of structurally distinct stores linked by control processes.

Information first enters a sensory register — modality-specific, high-capacity, and extremely brief. Visual input occupies iconic memory; auditory input occupies echoic memory. Most of what arrives there decays in a fraction of a second to a few seconds. Only the portion selected by attention is transferred into the short-term store.

The short-term store is capacity-limited and transient. Atkinson and Shiffrin already called it the subject’s “working memory”: the workspace in which control processes operate — rehearsal, coding, decision, and retrieval strategy. Unrehearsed contents drop out within seconds. Rehearsed or encoded contents enter the long-term store, which is relatively permanent and organized by meaning.

Two distinctions in the original model organize everything that follows.

First, stores are structural. Iconic memory is not a weaker version of working memory; working memory is not a weaker version of long-term memory. They differ in duration, capacity, code, and vulnerability.

Second, control processes are optional operations the system applies to those stores. Attention is the gate from sensory register to short-term store. Rehearsal and encoding are the gates from short-term store to long-term store. Retrieval is the operation that brings long-term content back into the working store so it can be used.

The tasks on this page are not generic “cognitive tests.” Each one loads a different store or a different control process.

Iconic memory is the visual sensory register. Sperling’s (1960) partial-report experiments showed that a brief visual display is available in near-complete form for only a few hundred milliseconds. The register is large; its duration is not. Unless attention selects an item from the icon, the trace is gone.

Echoic memory is the auditory counterpart. An auditory trace persists longer than an iconic one — on the order of seconds rather than milliseconds — which is why a question you were not quite attending to can still be “played back” if someone asks you immediately. Neisser named the store; Crowder and others measured its persistence.

These registers are not clinical curiosities. They are the first stage at which a processing failure can look like inattention, slowness, or poor memory when the actual problem is that the stimulus was never stably available for selection.

The Visual Threshold Serial Task (VTS) sits closest to this stage. It presents a rapid series of simple visual discriminations under time pressure, with minimal linguistic or mnemonic load. Because the items are brief, near-threshold, and sequential, VTS indexes whether iconic registration and sustained visual attention are intact enough to feed the rest of the system. Spared VTS with impaired span, Stroop, or lexical decision argues against a general sensory or vigilance failure. Impaired VTS with everything else slowed argues for a problem that begins before working memory and retrieval ever come into play.

Attention, in the modal model, is not a vague resource. It is the control process that determines which sensory contents enter the short-term store and which are lost. Once inside that store, a second attentional problem appears: the store is limited, and competing contents and competing responses must be managed.

Baddeley and Hitch (1974) fractionated Atkinson and Shiffrin’s short-term store into a working-memory system with specialized components:

• Central executive — attentional control: focus, division, switching, and inhibition of prepotent responses
• Phonological loop — brief storage and rehearsal of verbal-acoustic material
• Visuospatial sketchpad — brief storage and manipulation of visual and spatial material
• Episodic buffer (Baddeley, 2000) — a limited-capacity, multimodal interface that binds information from the slave systems and from long-term memory into a temporary episode

This is the architecture the performance tasks actually load.

Visual Digit Span presents digit sequences visually and requires immediate serial reproduction, with span length adapting to performance. Forward reproduction taxes passive storage — the sketchpad plus whatever phonological recoding the examinee applies. Longer and transformed sequences recruit the central executive. Span is not “memory” in the everyday sense; it is the measured capacity of the working store under immediate recall.

The Stroop Color-Word Task (Stroop, 1935) loads the central executive directly. Word reading is the prepotent, automatic response; color naming is the weaker, task-relevant response. The interference score is the cost of inhibiting the prepotent channel. That cost is an attentional-control cost, not a sensory or lexical one.

The Emotional Stroop substitutes affectively charged words for color words. Slowed color-naming on emotional items indexes attentional capture — the central executive losing the competition to salient affective content. In older language, this is conflict “reinvading” an autonomous function. In the operating model, it is an attentional-control failure under affective load. The two descriptions are compatible; only one of them specifies the stage.

The long-term store is not probed by asking whether someone “has a good memory.” It is probed by asking whether a stored representation can be accessed fast enough, and produced accurately enough, to meet a task demand.

Retrieval is access. A representation in the long-term store is activated, often automatically, often by spreading activation through a semantic network. The person may or may not produce anything yet.

Recall is production. The retrieved representation must be brought into the working store and output in the form the task requires — a digit sequence, a color name, a word/nonword decision.

The two dissociate. A person can retrieve a lexical item (it is “on the tip of the tongue”) and still fail to recall it. A person can recall a short digit string from the working store without ever having encoded it into long-term memory.

The Lexical Decision Tasks measure retrieval from the long-term lexical-semantic store under speeded recognition. Discriminating a word from a nonword in a few hundred milliseconds indexes a cascade: early visual feature extraction (iconic), orthographic encoding, phonological activation, lexical access, semantic retrieval, and response selection. Slowing can originate at any point in that cascade; the two LDT variants help locate it.

• The semantic-priming two-block LDT uses word pairs from the Semantic Priming Project pool. Facilitation from a related prime indexes automatic spread of activation through the semantic network — retrieval as network dynamics, not as deliberate search.
• The age-of-acquisition LDT manipulates when in development the words were learned. Early-acquired words are retrieved faster and more robustly. The AoA manipulation therefore probes whether the lexicon is organized along a normal developmental gradient, or whether retrieval is uniformly slowed regardless of when the item entered the store.

Visual Digit Span, by contrast, is immediate serial recall from working memory. It does not require that the digits have been encoded into long-term store. Converging a slow LDT with a spared span localizes the problem to long-term lexical retrieval rather than to the working store. The reverse pattern — intact lexical access, reduced span — localizes it to working memory.

The clinical value of the battery is the pattern, not any single score. Each instrument is a probe at a known point in the information-processing sequence.

Uniform slowing across VTS, span, Stroop, and LDT suggests a general processing-speed or attentional factor that begins early in the cascade. Selective impairment on Stroop with spared LDT and spared VTS points to executive inhibitory control, not to sensory registration or lexical retrieval. Selective slowing on emotional words with an intact classic Stroop points to attentional capture rather than a primary executive deficit. Slow LDT with spared span and spared VTS points to lexical-semantic retrieval. Reduced span with spared LDT and spared VTS points to the working store itself.

The older ego-psychological question — is the impairment primary (an apparatus that never functioned well) or secondary (an apparatus reinvaded by conflict, anxiety, or affective dysregulation) — is still the right clinical question. It is now asked inside a staged model. Affective load that consumes the central executive will look like a working-memory or attentional deficit. It should not look like an iconic-register failure, and it should not, by itself, dismantle semantic priming. Stage-specific sparing is what distinguishes a loaded control process from a broken store.