Week 11 - Individual Differences in Motor Control, Signal Detection Theory

Shams et al Sound induced flash illusions

• participants view a flash on a screen and hear beeps

• illusion: when two or more beeps are paired with a single flash, participants often perceive multiple flashes

• Findings:

  • as the number of beeps increases, the perceived number of flashes also increases

  • temporal binding window (TBW)

    • refers to the period during which multisensory stimuli are integrated

    • data suggest a wider TBW leads to stronger illusions

Individual exerperiences with the sound-induced flash illusion

• not everyone has the same TBW

  • some individual variations and differences in temporal binding

• variations exist but they may be too small at the neurophysiological level

temporal binding window (TBW)

• not consistent and differs among participants

  • can be identified using psychophysical methods

    • psychology + physics

How could we determine the perceptual threshold for someone?

• staircase method:

  • present pairs of stimuli at decreasing time intervals until participants report perceiving them as one or two separate stimuli

    • randomize intervals to prevent participant bias

• participants view visual stimuli presented in quick succession

• gradually adjust the speed of presentation to find the point where they can consistently detect them as separate events

• graph illustrating relationship between stimulus intensity and probability of detection or perception

  • threshold determination is typically at 50% detection probability

• Low intensity stimuli are often detected inconsistently

• different contexts can change the shape or position of the curve:

  • shift left: when the stimulus is easier to detect

  • shift right: when the stimulus is harder to detect

Signal detection theory - determining thresholds

• variability in detection thresholds can be determined mathematically

  • F(x)=(2πσ²)^-1/2exp[−(x−m)² / (2σ²​]

  • m = mean

  • σ = standard deviation

• described how values are distributed around a mean

Gaussian distribution

• how perceived intensities of stimuli are distributed in a population or experiment

• general bell curve with a peak at the mean M

• width of the curve is determined by the standard deviation reflecting variability

• greater the overlap between noise and signal distributions, the harder it is to distinguish signals

• noise-only distribution

• represents static or random noise present in the absence of a true signal

• when a real signal is introduced perceived intensity shifts to the right

• cuver overlaps with noise only curve showing uncertainty when identifying whether a stimulus is present or absent

• indiactes area where overlap occurs, representing uncertainty

Final explanation

• sensitivity is measured by the distance between the peaks of the noise and signal + noise distributions

• higher d = better detection

  • if the perceived intensity is greater than the criterion = stimulus detected

  • if perceived intensity is less than criterion = stimulus undetected

4 possible outcomes in SDT

• correct response (correct yes)

  • participant correclty identified it as present

• false alarm (incorrect yes)

  • no signal, participant iincorrectly identified signal

• correct rejection (correct no)

  • no signal, participant says there is no signal

• miss (incorrect no)

  • signal present, participant fails to detect signal

Fawcett - confusion matrix

• rows represent participants response classification (yes or no)

• columns represent actual presence of the stimulus (yes or no)

• graphs show how people set their decision criterion in signal detection tasks

• liberal:

  • The criterion shifts to the left, leading to more Hits but also more False Alarms.

  • Reflects a bias towards saying "Yes" to detect a signal even when uncertain.

• Conservative:

  • The criterion shifts to the right, leading to fewer False Alarms but also more Misses.

  • Reflects a bias towards saying "No" to avoid false positives.

Individual differences in sensory processing

• interested in variation sensory thresholds

• d-prime is a measure of sensitivity and estimates the standardized difference between the mean of two distribution

• steps to calculate d’:

  • see picutre

z-table

• used to convert proportions into Z-scores

  • e.g. if 85% of trials result in hits, located 0.85 in z-table to find corresponding z-score

  • represented how far a proportion is from the mean of a standard normal distribution

calculating d’

• hit rate (HR) and false alarm rate (FA) are used to calculate

  • measures an individual’s sensitivity in detecting a stimulus

• z-scores corresponding to HR and FA are substracted

why calculate d’?

• to be able to interpret the research on individual differences in sensory integration

• each person’s nervous system can use different criteria

  • can reflect a combination of individual differences in strategy and processing

Stevenson et al

• investigated whether variability in the TBW correlates with how individual perceive the McGurk effect

• McGurk effect:

  • phenomenon where conflicting auditory and visual stimuli are combined into a third perception

  • shows how sensory information is integrated or overridden by dominant modalities

  • e.g. hearing ba, while seeing ga, combined into da

stevenson et al results

• used psychophysics to determine the TBW for multisensory audiovisual stimuli

• found that duration of the right-side of the TBW was negatively correlated with likelihood of the perception of the McGurk effect

• individual with lower right TBWs are more likley to perceive the McGurk effect

Lebar et al

• examined participants brain response to visual stimuli during a mirror-reversed drawing task, comparing high and low performers

  • tasked with drawing shaped, with visual feedback presented normally or reveresed through a mirror

• Findings:

  • no differences in brain response to visual stimuli in mirror-reversed vs normal drawing conditions

  • high performers has increase visual cortex activation during drawing when compared to low performers

    • better integration of visual feedback with motor commands

    • heightened use of visual processsing to refine movements

Design of individual differences hypotheses

• how can we design an experiment to examin if high performers really do show increased brain acitvity?

  • used same techniques as lebar et al

    • training-studies: how training over time affects brain activation in low performers, are improvement in skill linked to changed in brain activity

    • between group differences studies: compare results of trained vs untrained, investigate if increased visual cortex activation is a trait of high performers or develops due to experience