Human Computer Interaction · Week 03

The Human as an Information Processor

Perceptual, cognitive and motor systems; vision, hearing and touch in UI design; sensory memory, STM (Miller's 7 ± 2, chunking) and LTM; cognitive load; attention; motor skills.

ITE37521 video · 11 min📄 Week 03 - Human Cognition, Memory & Ergonomics.pdf

▶Lecture videos

Click any timestamp — here or inside the notes — to jump the video to that moment. Your position is remembered.

Human as an Information Processing Machine

The human as input–processing–output, senses, memory and cognitive limits.

00:01Humans: the core component of HCI 01:39Designing interfaces humans can understand 02:33The human as an information processor: input and output 03:18Input channels: vision, hearing, touch 06:22Sensory memory → short-term → long-term memory 07:55Analysing sensory signals 08:51Recognition: matching against long-term memory (familiar faces)10:30Processing and responding

1The human as an information processing machine

HCI models the user as being like a computer, with three sub-systems:

Perceptual system (INPUT)

Cognitive system (PROCESSING)

Motor system (OUTPUT)

Figure 3.1: Human information processing sub-systems

Sub-system	Role	Computer analogy
Perceptual	Receives stimuli through the senses (eyes, ears, skin) and buffers them in sensory stores	Input devices + buffer
Cognitive	Processes, recognises, reasons, comparing signals with experience from long-term memory	CPU + RAM + disk
Motor	Turns decisions into muscle actions: clicking, typing, speech, head movement, steering	Output devices

Touching a burning hot surface

Input (perceptual): heat receptors in the fingertips detect intense heat.
Processing (cognitive): the brain recognises danger and tissue damage and issues an urgent motor command.
Output (motor): arm muscles contract instantly and the hand jerks away.
Post-cognitive evaluation: memory directs follow-up actions, like running cold water or applying ointment.

Cognitive process — Cognitive information flow: perception → working memory → long-term memory → response

2The senses in UI design

	👀 Vision	👂 Hearing	✋ Touch (haptic)
Nature	Directional: sharp foveal focus in the centre; peripheral vision detects movement and broad shapes	Non-directional (omni-directional): heard wherever the head is turned. Varies by pitch, timbre, volume	Needs direct physical contact: pressure, resistance, temperature, vibration
UI use	Critical content in the centre, secondary tools at the margins. The top-right corner is a known spot for alerts and profile toggles	Audio icons and chimes are great for urgent alerts	Personal and private: keyboard click-feel, phone vibration, VR haptic controllers
Caution	1 in 12 men and 1 in 200 women are colour-blind. Never rely on colour alone for status	Overuse overwhelms users and disturbs bystanders	—

3Human memory: three stores

Sensory memory (< 1 s)— attention →Short-term / working memory (4–5 items)— rehearsal →Long-term memory (permanent)

Store	What it is	Key facts
⚡ Sensory memory (perceptual store)	Ultra-short buffer of raw impressions: iconic (visual), echoic (auditory), haptic	Decays in under 1 second. Paying attention moves data into working memory
🧠 Short-term memory (STM / working memory)	Active scratchpad for immediate computation and decisions	Miller's Law: 7 ± 2 chunks in theory; realistically 4–5 items. Familiar items are kept far better than random strings
💾 Long-term memory (LTM)	Permanent store, built from STM through rehearsal and repetition	Episodic: serial, autobiographical events (graduation day). Semantic: a structured network of facts, concepts, relationships, skills

🧪 Try it — Short-term memory experiment

Press a button, memorise the sequence for 3 seconds, then type what you remember. Compare random letters with familiar chunks.

4Chunking & semantic networks

Chunking means grouping individual pieces of information into meaningful clusters. It greatly expands effective STM capacity without overloading it.

Meaningful words vs nonsense letters — Figure 3.3: Try to recall each set. Set 01 (real words) is far easier than Set 02 (meaningless letter groups): meaning helps memory

Phone numbers

0112650301 (10 separate digits) is hard to hold. 011 – 265 – 0301 is only 3 chunks, so it's easy.

🧪 Try it — Chunk it yourself

Any long string (phone, card, ID number)

Chunk size

Semantic memory: a network in the mind

LTM doesn't store isolated strings. Semantic memory is an interconnected associative network of nodes (concepts), properties and inheritance hierarchies.

Figure 3.5: Semantic network (concepts, properties and category links)

Example: "Animal" (breathes, moves) → "Bird" (has wings, can fly) → "Canary" (is yellow, sings). A canary inherits "can fly" from bird, so the fact doesn't need to be stored again.

5Cognitive load

Like a CPU with fixed RAM, the brain has finite cognitive capacity:

⚠️ Cognitive overload

Task complexity, visual clutter, stress and anxiety demand 100%+ of mental bandwidth, so error rates spike and performance drops sharply.

🥱 Cognitive underload

The task needs only about 30% of mental resources, so boredom sets in, leading to mind-wandering and attention drifting away.

🧪 Try it — Mental bandwidth meter

Task complexity Visual clutter Someone watching you

Why being watched makes you worse

When someone watches you work, performance anxiety and self-consciousness use up working memory. If the task itself needs a lot of resources, there isn't enough left, so you make errors in things you'd normally do easily. Try raising "Someone watching you" in the meter above.

6Attention: selective vs divided

	🎯 Selective attention	🔀 Divided attention (multitasking)
Meaning	All focus on one task stream, filtering out background distractions	Capacity split across several simultaneous tasks or rapid switching
Result	Better performance	Every context switch has a mental "adjustment penalty", so quality drops
UI goal	Clean, distraction-free layouts that protect focus	Minimise unnecessary interruptions and notifications

7Motor skills & physical ergonomics

Controls must fit human physical and biomechanical abilities:

Principle	Explanation
🎯 Hit target size and spacing	Make buttons large and well spaced to prevent accidental clicks. Don't put Close right next to Minimize
🏎️ Context-dependent precision	Touch precision is high sitting at a desk but drops when walking, running or driving on bumpy roads
📐 Fitts's Law	Time to reach a target depends on distance and target size. Large, close targets are faster and cause fewer motor slips (more in Week 9)

Exam focus

Typical questions:

"Explain the human as an information processor with an example" (the hot surface)
"Describe the three memory stores"
"What is Miller's law / chunking?"
"Differentiate episodic and semantic memory"
"Explain cognitive overload and underload"

Always link each one back to a UI design implication.

8Quick revision sheet

Concept	Remember it as
3 sub-systems	Perceptual (input) → Cognitive (processing) → Motor (output)
Vision	Directional; centre = critical; top-right = alerts; 1/12 men colour-blind
Hearing	Non-directional; good for alerts, but can annoy
Touch	Personal and private; vibration, key feel
Sensory memory	Iconic/echoic/haptic; < 1 s; attention moves it to STM
STM	7 ± 2 (Miller); realistically 4–5; chunking helps
LTM	Episodic (events) vs semantic (facts network)
Chunking	3–4 item groups, e.g. phone numbers
Cognitive load	Overload (100%+) → errors; underload (~30%) → boredom
Being watched	Anxiety uses working memory
Attention	Selective (better) vs divided (switching penalty)
Motor	Target size and spacing, context precision, Fitts's law

?Self-check quiz

11 questions on this week. Answer them all, then press Check my answers. Your best and latest scores are saved and shown on the subject page.

← PreviousInteraction Models, Ergonomics & Styles Next →Designing for Users — Tasks, Contexts of Use & User Perspectives