A Cognitive Paradigm to Investigate Interference in Working Memory by Distractions and Interruptions

Summary

A novel cognitive paradigm is developed to elucidate behavioral and neural correlates of interference by to-be-ignored distractors versus interference by to-be-attended interruptors during a working memory task. In this manuscript, several variants of this paradigm are detailed, and data obtained with this paradigm in younger/older adult participants is reviewed.

Abstract

Goal-directed behavior is often impaired by interference from the external environment, either in the form of distraction by irrelevant information that one attempts to ignore, or by interrupting information that demands attention as part of another (secondary) task goal. Both forms of external interference have been shown to detrimentally impact the ability to maintain information in working memory (WM). Emerging evidence suggests that these different types of external interference exert different effects on behavior and may be mediated by distinct neural mechanisms. Better characterizing the distinct neuro-behavioral impact of irrelevant distractions versus attended interruptions is essential for advancing an understanding of top-down attention, resolution of external interference, and how these abilities become degraded in healthy aging and in neuropsychiatric conditions. This manuscript describes a novel cognitive paradigm developed the Gazzaley lab that has now been modified into several distinct versions used to elucidate behavioral and neural correlates of interference, by to-be-ignored distractors versus to-be-attended interruptors. Details are provided on variants of this paradigm for investigating interference in visual and auditory modalities, at multiple levels of stimulus complexity, and with experimental timing optimized for electroencephalography (EEG) or functional magnetic resonance imaging (fMRI) studies. In addition, data from younger and older adult participants obtained using this paradigm is reviewed and discussed in the context of its relationship with the broader literatures on external interference and age-related neuro-behavioral changes in resolving interference in working memory.

Introduction

An extensive literature has demonstrated a detriment to the maintenance of information in working memory (WM) by interference from the external environment ^1-9. External interference can be classified into two general types; interference by irrelevant information one intends to ignore: distraction, and interfering information which demands attention as part of another (secondary) task goal: interruption. Studies comparing these types of external interference using a within-participant design enable assessment of the neuro-behavioral impact of goal-focused top-down attention in the processing and resolution of external interference.

Recently, the Gazzaley lab designed a paradigm that facilitates comparison of ‘to-be-attended’ interruptions and ‘to-be-ignored’ distractions that occur in the setting of a working memory task. Emerging evidence from this paradigm suggests that these different types of external interference exert distinct effects on behavior and have distinct underlying neural mechanisms ^2-5,10,11. This paradigm has revealed differences in external interference processing in normal aging ^2,3,4,10,11; though aging deficits in the context of interference are not always found ⁵; it has also distinguished mechanisms of interference by distractors versus interruptors using high-level visual stimulation of faces and scenes ^2,3,4,12, low-level visual motion of dot kinematograms ^5,10,11, and low-level auditory motion of frequency sweeps ⁵.

External Interference and Aging

External interference induces a detrimental impact on working memory throughout the lifespan, although older adults exhibit a more negative impact than younger adults ^2,3,13-18. Older adults also exhibit different patterns of neural activity compared to younger adults when attempting to resolve this interference ^3,4,17,21. However, some studies do not find evidence for such age-related behavioral ^5,19,20 or neural ⁵ differences with interference.

Interestingly, the impact of aging on resolving interference seems to differ by sensory modality, although this issue remains unresolved at present. Visual intrasensory interference has been widely shown to exhibit age-related decline (summarized in an extensive review ²²). In contrast, many experiments suggest no age-related deficits during intra-sensory auditory interference ^19,22-25, while other studies demonstrate significant age-related increases in auditory distractibility ^19,22,26-32. In addition, the salience of interfering stimuli (congruent or incongruent between the cue and probe stimuli) ², and stimulus complexity (high or low processing load) ⁵ may interact with interference processing and its differences across task goals and age.

The paradigm described here supplements the aging interference literature by probing the mechanisms of top-down attention (in the form of task goals) and resolution of external interfering stimuli. Evidence from the visual face & scene version of this paradigm indicates an interaction between aging and interference type, with older adults demonstrating even greater vulnerability to attended interruptors relative to ignored distractors ^3,4. Characterizing the behavioral and neural differences between these types of interference are important to understand how cognitive control abilities change with aging.

Why do older adults show exacerbated deficits in resolving to-be-attended interruptors? Are older adults impaired by excessive processing of interruptors when they are presented, or by an inability to re-activate representations of the primary goal-relevant stimuli after interruptions, or by prolonged processing of interruptors after they are no longer present or relevant ³³? To address these questions, the current paradigm’s design allows for comparison of neural activity at time-points before, during, and after different types of interference. For instance, by comparing neural activity elicited by ignored distraction versus activity during attended interruptions, one can ascertain the specific impact of top-down attention on resolution of interference in working memory.

Several studies have implemented multiple variants of this interference paradigm to understand the neural correlates of the different types of external interference both at high spatial and temporal resolution using functional magnetic resonance imaging (fMRI) and electroencephalography (EEG), respectively. This paradigm has also been used to clarify important distinctions between interference in the visual and auditory domains, as well as the impact of stimulus complexity and congruence on interference. Here, the paradigm variants are described in detail.

Protocol

The steps below enumerate how to execute this novel cognitive paradigm designed to elucidate the neuro-behavioral aspects of external interference on delayed recognition working memory, with variations optimized for pairing with EEG or fMRI. Prior to beginning data collection, complete all necessary human-participants research approvals through the appropriate Institutional Review Board and/or human participants review committee.

1. Preparation

1. Download and install experiment presentation software such as E-Prime, Presentation, or PsychoPy, as per manufacturer’s instructions, onto a dedicated stimulus presentation computer.
2. Prepare an appropriate keypad for experimental responses. Add “YES” and “NO” labels to two adjacent keys (Figure 1).
   NOTE: For versions of this experiment utilizing MRI, use an MR-compatible keypad.
3. For auditory versions of this paradigm, prepare headphones appropriate for testing modality (ie: EEG or MR-compatible, if necessary), as per manufacturer’s instructions, and adjust sound level for presentation at 65 decibels (dB) sound pressure level (SPL), which is a comfortable level for normal hearing individuals.
4. For experiments with older adults, conduct preliminary neuropsychological and sensory screenings such as vision and hearing to select an appropriately screened study population.
   1. Neuropscyhological screening
      1. Create a neuropsychological assessment battery to screen for cognitive impairment in older adults. Administer tests by paper-and-pencil, or adapt a battery for testing on a computer.
         NOTE: Tests may include the Mini Mental Status Exam (MMSE) ³⁵ , Global Deterioration Score (GDS) ³⁶, California Verbal Learning Test (CLVT) ³⁷, Digit Span ^38,39, Symbol Span ⁴⁰, Letter-Number Sequencing ⁴¹, Delis-Kaplan Executive Function System (D-KEFS) – Trail Making Test ⁴², Controlled Word Association Test (COWAT) ^{43, 44}.
      2. Administer this battery to all prospective adult participants. Score all tests per their respective scoring guidelines.
      3. If recruiting for healthy older adults, exclude prospective participants with scores greater than two standard deviations below the population mean, or per custom exclusion criterion.
   2. Vision screening
      1. For visual experiments, screen for normal or corrected-to-normal vision using a preliminary questionnaire asking whether participants have normal or corrected-to-normal vision.
      2. To follow up, conduct a Snellen chart vision test, and exclude participants without normal or corrected-to-normal (20/20 or greater) vision.
   3. For auditory experiments, screen for normal hearing:
      1. In a preliminary questionnaire, ask whether participants have normal or corrected-to-normal hearing, and exclude those who do not.
      2. To follow up, obtain an objective measurement of hearing sensitivity. Conduct an in-lab audiometric assessment with one of several methods:
         1. Utilize a hearing loss screening test application such as ‘uHear’. Using this application’s auto-calculated results, exclude subjects with hearing sensitivity outside of the ‘normal hearing’ range.
         2. Assess audiometric thresholds in the 250 - 6,000 Hz frequency range in both ears by the method of ascending and descending limits. Individuals with mean audiometric thresholds greater than 50 dB at any test frequency in either ear, signifying moderate hearing loss, should be excluded

2. Experimental Design

1. Administer a delayed recognition working memory task under three distinct interference conditions (and a fourth baseline condition for neural experiments) in a block design (see also Figure 2 and Table 1). Repeat each condition twice, in counterbalanced order (a balanced Latin square design is recommended). Note that experimental timing and number of trials vary between paradigm variants; utilize the parameters detailed in Table 1.
2. Ignore Distracting Stimulus Condition (DS):
   1. Display a prompt instructing participant to remember the cue stimulus and ignore the distracting stimulus while continuing to maintain a representation of the cue stimulus. Instruct the participant to respond “YES” if the probe stimulus matches the cue stimulus or “NO” if the probe does not match the stimulus.
   2. Present the cue stimulus, immediately followed by a short delay (Delay 1).
   3. Display an interfering ‘distractor’ stimulus, immediately followed by a second short delay (Delay 2).
      NOTE: The participant does not need to (and should not) interact with the distractor stimulus.
   4. Present a probe stimulus and collect responses.
3. Attend to Interrupting Stimulus (Secondary Task) Condition (IS):
   1. Display a prompt instructing participant to remember the cue stimulus and complete a secondary task using the interfering stimulus that appears thereafter. Display instructions to complete the secondary task as follows, “press a button only if the interrupting stimulus matches a set of discrimination criteria”. Instruct the participant to respond “YES” if the probe stimulus matches the cue stimulus or “NO” if the probe does not match the stimulus.
      NOTE: The discrimination criteria are distinct for each paradigm variant and described in the next section.
   2. Present the cue stimulus, immediately followed by a short delay (Delay 1).) Present an interfering ‘interruptor’ stimulus and collect responses for the secondary (discrimination) task. Following, present a second short delay (Delay 2).
      NOTE: Completing the secondary task requires attention to the ‘interruptor’.
   3. Present a probe stimulus and collect responses.
      NOTE: Ten percent of trials are catch trials in which the interruptor matches the discrimination criteria; add additional trials (10%) to this block to compensate for the discarded trials. Exclude all catch trials from neural analysis due to the confounding motor response.
4. No Interfering Stimulus Condition (NI):
   1. Display a prompt instructing the participant to remember the cue stimulus and keep it in mind. Instruct the participant to respond “YES” if the probe stimulus matches the cue stimulus or “NO” if the probe does not match the stimulus.
   2. Present the cue stimulus, immediately followed by a delay. Display a central fixation cross on a blank screen during the delay.
   3. Present a probe stimulus and collect responses.
5. Baseline/Passive View (or Listen) Condition (only for neural experiments) (PV/PL)
   1. Include a passive-view/listen condition during neuroimaging tasks to enable calculation of ‘enhancement’ and ‘suppression’ of neural activity during IS/DS conditions relative to baseline activity when participants passively view (/listen to) the working memory and interfering stimuli, free from task goals. (See Table 2.)
   2. Display a prompt instructing participant to passively view (/listen to) all visual (/auditory) task stimuli. Display instructions to complete the simple discrimination task.
      1. For visual tasks, instruct the participant to press a button corresponding to the direction of a displayed arrow (left or right).
      2. For auditory tasks, instruct the participant to press a button corresponding to the frequency range of an easily discriminable high (2 kHz) or low (0.5 kHz) frequency sound sweep (high or low).
   3. Sequentially present or display the cue stimulus, Delay 1, interfering stimulus, and Delay 2.
   4. Present an arrow (visual) or sound sweep (auditory) in place of the probe stimulus and collect responses as the participant completes the simple discrimination task (described above).

3. Stimuli

1. General Preparation of Stimuli

1. Select a set of stimuli from the categories described below (see also Figure 2 and Table 1).
2. Carefully decide whether to pair primary working memory task stimuli with thematically congruent or incongruent interfering stimuli (see NOTE below).
3. Ensure that all images are sized or re-sized to 225 pixels wide and 300 pixels tall (14 x 18 cm).
4. Present images foveally, subtending 3 degrees of visual angle from fixation.
   NOTE: For fMRI experiments, use interference stimuli incongruent with the primary working memory task stimuli, for example, face interference during scene working memory or vice versa. To precisely localize face and scene specific sensory cortical regions, apply a fMRI localizer task prior to the working memory experiment. Then, during the interference paradigm, use these scene and face selective cortical regions to simultaneously parse neural activity dynamics to the working memory cue stimuli (e.g., scenes) and to the incongruent interference stimuli. (e.g., faces)

2. High-level Visual Stimuli

1. For face stimuli, prepare several hundred Cue/Probe Face stimuli from gray-scale photos of male and female faces, with neutral expression, across a large adult age range. Remove hair and ears digitally, and apply a blur across the contours of the face.
2. For scene stimuli, prepare several hundred Cue/Probe Scene stimuli from gray-scale photos of natural scenes.
3. After Delay 1, present a interfering stimulus consisting of a scene or face. On 90% of trials, present a face that is not ‘male AND aged over 40 years old’; on the other 10% of trials, present a face that is male and aged over 40 years old.
4. For “Attend to Interruption” condition, instruct participants to complete the following secondary task using the interfering stimulus (presented between the cue and the probe). Ask the participant to respond “YES” if interrupting face is male and aged over 40 years old.

3.  Low-level Visual Motion Stimuli

1. Create Cue/Probe stimuli of a circular aperture containing 290 spatially-random grey scale dots (0.08 degrees x 0.08 degrees each) that subtend 8 degrees of visual angle at a 75 cm viewing distance, centered at the fovea.
2. Display moving dots with 100% motion coherence at an oblique angle of 10 degrees per sec, at one of 12 different directions of motion (3 in each sector).
3. Use an adaptive staircase thresholding procedure (2 degree increments) to establish a visual discrimination value yielding just under 100% accuracy, such that the discrimination threshold is reached upon the first error trial.
4. After Delay 1, present an interfering stimulus consisting of dots in counter-clockwise circular motion. Render this motion at a ‘normal’ speed (10 degrees per sec) on 90% of trials, and fast on the other 10% of trials.
5. In the Attend to Interruption condition, instruct participants to complete the following secondary task: respond “YES” if interrupting swirl is fast.

4.  Low-level Auditory Motion Stimuli

1. Create Cue/Probe Stimuli of sound motion sweeps across a frequency range with mid-frequencies randomly chosen between 900 and 1,100 Hz. Construct the sound motion sweep frequencies to start at ± 0.5 octaves from the mid-frequency and end at ± 0.5 octaves from the mid-frequency.
2. Present an equal portion of ‘up’ (starting at -0.5 and ending at +0.5 octaves) and ‘down’ (starting at +0.5 and ending at -0.5 octaves) motion sweep stimuli.
3. Adjust the volume to comfortable hearing level of 65 dB SPL.
4. Thresholding: use an adaptive Zest procedure to establish auditory discrimination accuracy at 85% correct performance.
5. After Delay 1, present an interfering stimulus consisting of a single tone. Play a tone of frequency 2 kHz on 90% of trials, and a tone of 2.3 kHz on the other 10% of trials.
6. In the Attend to Interruption condition, instruct participants to complete the following secondary task: respond if interrupting tone is a higher frequency cue (2.3 kHz).

5. Probe Stimuli

1. For all WM tasks, ensure that 50% of probe stimuli match the cue.
2. In the low-level motion tasks with thresholded discrimination levels 5,10,11, set 50% of the probe stimuli, which do not match the cue, to differ from the cue by the absolute value of the participant’s thresholded stimulus discrimination level.
   NOTE: For instance, if thresholding establishes a participant’s visual discrimination level to be 10 degrees, pair a visual motion cue moving at 45 degrees with a probe moving at either 45 degrees (match on 50% trials) or 45 ± 10 degrees (35 or 55 degrees; each non-matches on 50% trials).

4. Comparing Interference Conditions

1. Use statistical software, such as SPSS, to compare behavioral performance and neural activity at important time-points before, during, and after different types of interference.
   NOTE: Several manuals online provide step-by-step instructions and screenshots describing how to use and run simple statistical analyses in SPSS.
   1. Calculate the impact of distractions versus interruptions on behavioral performance by contrasting working memory accuracies and response times during the interference conditions relative to the performance during the no interference condition (Figure 4). For instance, paired t-tests can be used to compare accuracy or RT between any two interference (or baseline) conditions.
      NOTE: Prior to t-test comparisons between two specific task conditions, a repeated measures ANOVA is recommended to compare across all working memory conditions in the paradigm .
   2. For neuro-imaging studies, pre-process and process the data according to the appropriate pipline for the modality and measures of interest.
      1. For EEG studies, process EEG data with EEGLAB or the software package of choice, using software’s instructions and recommended processing stream.
      2. For fMRI studies, process fMRI data with the software package of choice (such as AFNI, SPM, FSL, etc.), using software’s instructions and recommended processing stream.
   3. To assess neural activity modulations as a consequence of interference during working memory, statistically contrast neural data in these conditions to neural activity during passive view(/listen) conditions, thus controlling for basic perceptual processing (Figure 4).
      1. Calculate the Measurements such that a positive value always indicates greater enhancement above baseline or greater suppression below baseline. For P100, calculate neural suppression by subtracting quantified neural activity to the distracting stimulus (DS) from that evoked by the passively viewed stimulus (PV) (i.e: PV - DS). Calculate enhancement in fMRI by subtracting quantified BOLD activity to the baseline passively viewed stimulus from that evoked by the interrupting stimulus (IS) (i.e: IS – PV).
   4. Statistically compare neural modulations elicited by ignored distractions versus activity during attended interruptions to begin to ascertain the specific impact of top-down attention on the resolution of different types of interference in working memory.

Results

This interference paradigm has enabled generation of important findings regarding the distinct behavioral impact and neural mechanisms of distraction and interruption on working memory in younger and older adults (see Table 2 for summary).

Behavior. Behaviorally, in line with the existing literature, interruption consistently imparts a greater detrimental impact versus distraction on working memory performance ^{2-5, 10,11,12}. Older adults exhibit eve...

Discussion

A novel cognitive paradigm has shown efficacy in investigating working memory interference by distractions and interruptions. This paradigm and its several variants, extending its use across sensory modalities, stimulus complexity levels, and imaging methods, are detailed.

Before beginning the experiment, pre-screen all participants to ensure appropriate cognitive and perceptual abilities. For experiments using low-level perceptual stimuli, administer an adaptive thresholding procedure to cali...

Materials

Name	Company	Catalog Number	Comments
Computer for stimulus presentation	Dell	Optiplex GX620	hardware/software requirements will vary based on stimulus presentation software
Cathody Ray Tube (CRT) monitor	ViewSonic	G220fb	21"; recommended due to its superior latency relative to that of LCD monitors in displaying visual stimuli; chair should be positioned 75 cm away
E-Prime software	Psychology Software Tools, Inc.	E-Prime 2.0 Standard	a different experimental presentation software can be used in place of E-Prime (e.g. Presentation (Neurobehavioral Systems), or PsychoPy (open-source); E-Prime and Presentation are compatible with Microsoft Windows, PsychoPy is compatible with Microsoft Windows, Mac OS X, and Linux)
Keyboard/response pad for Behavioral or EEG experiments	Keyboard: Razer; Response Pad: Cedrus	Keyboard: BlackWidow Ultimate; Response Pad: RB-830	any standard computer keyboard is acceptable, though response pads may offer more precise timing (ie: Cedrus RB-830 guarantees 1 ms resolution)
Keyboard/response pad for MRI experiments	Curdes	Package 904	ensure that keypad is MR-compatible
Headphones (for auditory behavioral experiments)	Koss	UR29
EEG-compatible Headphones (for auditory EEG experiments)	Etymotic	ER3-50; ER3-21; ER3-14A
MRI-compatible Headphones (for auditory MR experiments)	Etymotic	SD-AU-EAER30