Week 2: Methods and Sensorimotor Capacities

Gaze-based methods: strengths and weaknesses

Main tool for studying infants (preferential looking, habituation, familiarity/novelty, VoE)

Strengths: works with very young infants; easy to run on a screen

Weaknesses: gaze itself develops with age; the link between gaze and cognition is weak (many things drive looking); needs many control experiments

Preferential looking

Method: Two stimuli side by side. Measure which one the infant looks at longer.

What it tells us: longer looking at one = the infant can tell them apart (or prefers one).

Habituation

Method: Show one stimulus over and over until looking time drops (infant gets "bored" as it encodes it). Then show a new one.

Habituation criterion: each baby goes at their own pace; habituated once looking drops to ~half of their first few trials.

Dishabituation: if looking recovers to the new stimulus → infant sees it as different.

Also measures processing speed and memory.

Hunter and Ames 1988

Familiarity → novelty preference model

Early in learning, infants prefer the FAMILIAR stimulus. After they fully encode it, they prefer the NOVEL one.

Explains why some studies find familiarity preferences and others novelty preferences. Either one shows discrimination.

Violation of expectation (VoE)

Method: Show a possible event and an impossible event, usually after familiarizing infants with the setup.

Longer looking at the impossible event = surprise → the infant expected something else.

Controls: an impossible scene may just be more interesting, so studies measure babies' baseline preference for each final scene.

High-amplitude sucking (HAS)

Age: newborns

Strengths: natural behavior from birth. Infants suck more for things they like and can learn to suck more or less to get a reward.

Weaknesses: long, complicated sessions (need baseline sucking rate, training, an awake baby). Sucking declines with age.

Reaching (method)

Age: from ~5 months

Strengths: reaching shows preference

Weaknesses: can't use with screens; binary (reach or not), no measure of intensity

Language and touchscreen tasks

Age: ~2–3 years

Strengths: flexible tasks; 2-year-olds can use iPads

Weaknesses: children differ in tech experience; once they talk, they may say what they think the experimenter wants to hear

Physiology (heart rate, sweat)

Age: whole lifespan

Strengths: easy and naturalistic

Weaknesses: only works for tasks that evoke emotion

Neuroimaging (EEG, MEG, fNIRS, fMRI)

EEG, MEG, fNIRS (blood flow), fMRI (good spatial resolution)

Age: whole lifespan

Strengths: no behavior needed; can rely on the brain alone

Weaknesses: expensive, technically demanding, high attrition, noisy signal that doesn't always reflect cognition

Basirat, Dehaene and Dehaene-Lambertz 2014

Local-global sound paradigm (EEG)

Question: Do infants track short-term patterns, longer-term patterns, or both?

Method: Infants hear tone sequences while EEG records. Local deviant: a surprising high tone at the end (breaks the short pattern). Global deviant: a surprising low tone at the end (breaks the longer pattern of 4 low + 1 high).

Result: The brain responds differently to each kind of deviant, showing which patterns infants track.

Why it matters: neural measures show expectations without any behavior.

Theme: reducing demands on behavior

Longitudinal design

Test the SAME group many times over development

Strength: within-participant, so it controls for baseline differences between individuals

Weakness: attrition (participants drop out, data lost)

Cross-sectional design

Test DIFFERENT age groups at the same time

Strength: fast; low attrition

Weakness: cohort effects; baseline differences between groups

Experimental vs. observational designs

Experimental: manipulate an independent variable (IV) and measure a dependent variable (DV). Strength: allows causal claims. Weakness: needs strong controls.

Observational/correlational: observe without manipulating. Strength: useful when experiments aren't possible. Weakness: causation is hard to establish.

Samples and statistical power

Convenience samples: most studies use them; biased and not representative

WEIRD samples: Western, Educated, Industrialized, Rich, Democratic

Statistical power: rises with larger samples (>~20 per group), which is hard to reach with infants

Sensation vs. perception

Sensation: raw input to the sense organs (e.g., wavelengths of light hitting the retina)

Perception: interpretation of that input (e.g., seeing a color)

Measuring acuity with stripes

Question: How sharp is infant vision (spatial acuity)?

Method: Preferential looking: striped pattern on one side, blank gray patch on the other. Infants prefer stripes if they can see them. Stripes get thinner until the infant no longer prefers them (acuity in arc minutes). Second version: EEG as stripes get finer.

Result: Finest stripes the infant still prefers = acuity. EEG gives SHARPER acuity estimates than looking.

Claim: Methods that rely less on behavior give more accurate results.

Theme: reducing demands on behavior

Why acuity improves

Photoreceptors in the retina mature

Feedback connections in the visual system increase and get faster with development

Testing color with habituation

Question: Can infants see color (hue), not just brightness?

Method: Habituate to one color, then show a similar color or a clearly different color.

Result: If looking recovers for the different color, the infant can tell hues apart.

Claim: Young babies see luminance (brightness) but can't yet discriminate hue.

Visual cliff (Eleanor Gibson)

Question: Do infants perceive depth?

Method: Locomotor behavior. Infant on a glass surface; under one side the floor looks like a sudden drop. Mom calls from the other side. Does the infant cross the "cliff"?

Result: Crawling infants avoid the deep side.

Claim: Infants perceive depth, using binocular cues (disparity, convergence) and monocular cues (occlusion, relative size, vanishing point).

Crawling vs. walking on the visual cliff

A crawling infant may refuse to go over the visual cliff, yet the same infant may go over it soon after learning to WALK.

Claim: Locomotion knowledge doesn't transfer to a new form of movement.

Theme: development does not equate to improvement

Hearing development

Hearing is mature earlier than vision (ear formed by birth; retina still developing)

Newborns prefer speech to non-speech

Newborns localize sounds (look toward the source)

Fetal hearing (DeCasper and Spence)

Mothers read a story aloud during pregnancy. After birth, babies preferred mom reading that story over a different voice reading a different story.

Claim: Hearing works before birth.

(Full study in Week 5)

Smell, taste and pain

Newborns recognize mom by smell before sight

Taste experience in the womb shapes later food preferences

Newborns are MORE sensitive to pain than adults

Newborn reflexes

Humans need a "fourth trimester" (long development after birth); newborns move mostly through reflexes.

Rooting: stroke the cheek → turns toward it and sucks

Sucking: sucks even without feeding (non-nutritive)

Grasp: grips an object placed in the hand

Stepping: held above ground → makes stepping movements

Moro: at a loud sound → throws arms out

Eye movement control (order)

1. Sticky fixation: can't look away from a stimulus

2. Saccades: eyes jump between points

3. Smooth pursuit: smoothly follows a moving target

4. Goal-directed looking: looks for or at something on purpose

Motor milestones (average ages)

Lifts head: 1.5 months

Chest up on stomach: 2 months

Rolls over: 4.5 months

Sits without support: 6 months

Crawls: 8 months

Stands without support: 11 months

Walks: 12 months

Variability and culture in motor development

Variability: less tummy time slows motor development; bulky diapers disrupt walking; low-birthweight and premature infants reach milestones later

Culture: cultures that exercise infants (tossing, making them stand) → faster motor development; cultures that hold infants in place or keep them on their backs → slower

Much of this research comes from WEIRD samples

Theme: studying variability enhances understand

Moser: local-global paradigm (newborns and fetuses)

Origin: coma patients hear tone sequences. Local deviant = surprising high tone after 4 low tones. Global deviant = 5 low tones, breaking the repeated 4-low + 1-high pattern. Patients whose brains showed the global-deviant response later recovered; those without it didn't -> signature of consciousness.
Moser's question: do newborns and fetuses show the global-deviant response?
Method: EEG in newborns (awake vs. asleep); MEG through mom's abdomen for fetuses >35 weeks.
Result: AWAKE newborns and active fetuses showed it; SLEEPING ones did not.
Claim: neural measures reveal processing with no behavior (doesn't tell us what the experience is like).
Theme: reducing demands on behavior

Fantz stripes and VEP: measuring acuity

Question: how sharp is infant vision (spatial acuity)?
Method 1 (Robert Fantz, 1960s): forced-choice preferential looking. Stripes (grating) on one side, gray patch on the other. Stripes get finer in a staircase until looking is at chance (50%).
Method 2: sweep VEP (EEG). One eye covered, gratings get finer over ~5 min, electrodes record visual cortex. No looking needed.
Result: newborns are at chance for 20-arc-minute stripes (a matchstick at arm's length). VEP gives HIGHER acuity than looking at every age, with tiny confidence intervals.
Claim: methods that rely less on behavior give more accurate results; gaze itself is immature.
Theme: reducing demands on behavior

Teller acuity cards

Davida Teller's fast version of forced-choice preferential looking (FPL) for measuring infant acuity.
Each card has a grating (stripes) on one side and plain gray of equal brightness on the other.
An observer who can't see which side has the stripes watches the baby's looking and judges where the grating is.
Cards go from wide to fine stripes; the finest grating the observer can still locate from the baby's looking = acuity.
(Not named in the recorded lecture; check the slides.)

Arc minutes / visual angle

60 arc minutes = 1 degree of visual angle
Thumbnail at arm's length ~ 1 to 1.5 degrees
20 arc minutes ~ a matchstick at arm's length (newborns can't resolve it)

Color: hue and color categories (Bornstein 1976)

Newborns see luminance (brightness) but are poor at hue; hue discrimination emerges over the first few months.
Categories: habituate 4-month-olds to a blue. Test two colors equally far in wavelength: another blue (same category) vs. a green (across the boundary). Infants stay habituated to the blue and recover to the green. Same for green/yellow, yellow/red.
Claim: by 4 months infants carve color into adult-like categories.
Nativist: built in. Empiricist (Locke): hundreds of hours of experience; categories differ across cultures.
Theme: nature and nurture
(Bornstein, Kessen and Weiskopf 1976 - not named in lecture)

Visual cliff (Gibson and Walk)

Question: do infants perceive depth?
Method: crawling infant (6-13 months) on glass; one side looks like a sudden drop; mom calls from the far side. Kittens and chicks tested too.
Result: crawling infants (and kittens, chicks) avoid the deep side.
Their claim: depth perception is innate or emerges around when locomotion starts.
Problem: the task needs crawling. Lower-demand tasks find depth perception earlier (binocular ~4 mo, pictorial ~6-7 mo).

Binocular vs. monocular depth cues

Binocular (needs both eyes): retinal disparity (each eye sees a slightly different image), convergence
Monocular / pictorial (one eye is enough): occlusion, relative size, linear perspective (vanishing point), texture gradient

Which depth cues emerge first: binocular or monocular?

BINOCULAR first: ~4 months
MONOCULAR (pictorial) later: ~6-7 months (5-month-olds don't use them)
Both appear before/around crawling, so the visual cliff can't tell us when depth perception starts.
Note: textbooks say kinetic (motion/looming) cues work even earlier (first month), but the lecture's monocular cues = pictorial cues.

Ponzo duck illusion (Yonas)

Question: do infants use pictorial (monocular) depth cues?
Method: two identical ducks on a flat picture of a grid receding into the distance (linear perspective). Adults see the 'farther' duck as bigger/farther. Measure which duck the infant REACHES for (infants reach for the nearer object).
Result: 7-month-olds (lecture: ~6 mo) reach for the duck that looks closer; 5-month-olds don't.
Claim: pictorial depth cues emerge ~6-7 months, after binocular cues.
(Yonas and colleagues - not named in lecture)

DeCasper and Spence 1986 (The Cat in the Hat)

Question: do fetuses hear and remember speech?
Method: moms read The Cat in the Hat (or a same-rhythm story with different words) aloud daily in the last weeks of pregnancy. A few days after birth: HAS task; sucking faster or slower (counterbalanced) played mom reading that story vs. another mom reading a different story.
Result: newborns sucked to hear the familiar story.
Claim: hearing works before birth; fetuses learn and remember speech.
(Don't put headphones on a pregnant belly - can damage hearing.)

Newborn cry melody (French vs. German)

Newborns' first cries carry the melody (accent) of the language heard in the womb: French and German newborns cry differently.
Supports DeCasper and Spence: fetuses learn from speech before birth.
(Mampe and colleagues 2009 - not named in lecture)

Fetal heart rate and voices

Question: can fetuses tell mom's voice from a stranger's?
Method: play recordings of mom's or a stranger's voice to fetuses; record fetal heart rate.
Result: heart rate RISES for mom's voice, FALLS for a stranger's.
Claim: fetuses distinguish speakers before birth. Limit: heart rate shows a difference, not the experience.
(Kisilevsky and colleagues 2003 - not named in lecture)

Infant pain

Newborns are MORE sensitive to pain than adults: they lack the inhibitory pathways that dampen pain
Until the late 1980s many believed newborns felt no pain
Distraction (engaging another sense) reduces pain during shots

Fourth trimester

Humans are born very immature; much development happens after birth.
Why: bipedalism limits hip width, which limits head size at birth.
Contrast: a foal stands and walks ~30 minutes after birth; humans take ~12 months.

Eye movement control (order and ages)

Eye movements = first movements under voluntary control
1. Sticky fixation: can't disengage from a stimulus until ~2 months (1-month-olds may stare up to 20 min)
2. Saccades: newborns undershoot (hypometric, several small jumps); mostly accurate ~6 months; adult-like ~8 years
3. Smooth pursuit: little before 2 months; clear by ~5 months
4. Goal-directed (anticipatory) looking: by ~4 months, learn a dull cue predicts an interesting picture on the other side and look there before it appears

Why walk instead of crawl?

Faster
New walkers fall about as often as crawlers (~39 falls/hour, usually no fussing)
Richer view of the world (crawlers mostly see the floor)
Travel farther
Hands free to carry objects and bring them to people

Adolph: adjustable drop-offs

Question: is knowing which drop-offs are safe innate, and does it transfer from crawling to walking?
Method: Karen Adolph's adjustable apparatus: drop-offs from a small step up to a 90 cm cliff; a 'baby catcher' stands by. Experienced crawlers and novice walkers (12 months), experienced walkers (18 months).
Result: experienced crawlers judge within ~1 cm and refuse unsafe drops. Novice walkers walk straight off, even 90 cm. After months of walking, they again judge within ~1 cm and scoot/back down.
Claim: no innate fear of heights; learning doesn't transfer to a new posture.
Theme: development does not equate to improvement

Clearfield: search skill and walking

Question: does spatial skill learned while crawling carry over to walking?
Method: infants search for hidden toys; compare weeks of crawling experience with success, then retest after they start walking.
Result: more crawling experience -> better search (positive correlation). New walkers get WORSE.
Claim: what infants learn about space is tied to their current way of moving.
Theme: development does not equate to improvement

Culture: restricting vs. exercising infants

Restricting: Tajikistan gahvora cradle (~20 hrs/day, birth to ~2 yrs); northeastern China sandbags -> slower milestones
Exercising: parts of Africa, the Caribbean, etc. (lifting, tossing, practice sitting/standing) -> faster milestones; trained infants can skip crawling
Don't judge as good/bad parenting: practices protect infants in hazardous settings and come with compensating practices
Theme: studying variability enhances understanding

Cameroon sitting at 5 months (Karasik 2015)

Question: does culture change when milestones appear?
Method: minutes 5-month-olds can sit independently, across six cultures.
Result: most 5-month-olds barely sit; in Cameroon many sit alone up to ~20 minutes.
Claim: motor development depends on experience and practice, not just a built-in schedule; earlier sitting has cascading benefits (upright view of social life).
(Karasik and colleagues - not named in lecture)

Week 2 examples: reducing demands on behavior

VEP (EEG) shows higher acuity than looking
Looking/reaching show depth perception (binocular ~4 mo) before crawling allows the visual cliff
Moser: brain responses in newborns and fetuses with no behavior

Week 2 examples: development does not equate to improvement

Adolph: crawlers avoid unsafe drop-offs, new walkers walk off them
Clearfield: search skill drops when infants start walk