Specific Learning Disabilities: The Role of Working Memory and - - PowerPoint PPT Presentation

Lisa Archibald, PhD Western University, Canada

Cogmed Conference 2014

Specific Learning Disabilities: The Role of Working Memory and Other Domain-specific Deficits

Specific Learning Disabilities

• Appear to have typical potential to learn
• Adequate experiences & educational
• pportunities
• But, difficulty learning in one or more areas

Specific Language Impairment Dyslexia Dyscalculia

Specific Learning Disabilities

Specific Learning Disabilities

• Heterogeneity

– Different underlying cognitive impairments?

• Domain-general

– Memory, executive functioning, etc.

• Domain-specific

– Phonological processing, magnitude processing

Systems Supporting Learning

Knowledge base (LTM) *largely mediated by language Performance- based factors *working memory

Working Memory & Learning

• Actively monitors & manipulates information

in current focus of attention

– New information – Needed information – Transformed / derived information

• Early in the learning process
• Supports development of the knowledge base

Working Memory and Language

• WM may have specific impact on language

learning

– Language is delivered in a rapid code – Language is largely delivered via acoustic waveforms

• Vanish rapidly
• Time-dependent

Bishop, 1992; Kail, 1994

Working Memory and Literacy

• Reading dependent on decoding

– Arbitrary connection between speech sounds and letters

• Effortful decoding places high demands on

WM

– Retention for comprehension

Working Memory and Numeracy

• Math relies on connections between numbers

and symbols

– Numerical symbols are arbitrary

• Operations require retention of intermediate

products

Language & Learning

• Stores of knowledge based on human verbal

code for communication

• Learned over multiple trials
• Fairly automatic once learned
• Supports learning of related concepts

(networks of knowledge)

• Increasing importance over learning process

Language Scaffolds Language Learning

• Verbal skills support language learning

– Familiar word types easier to remember

• E.g. daevacheenoitag vs. trumpetine

– New grammatical forms that fit existing rules

• E.g., wuffed; tweet/twat

– Sentence comprehension with known vocabulary/context

Language and Literacy

• Word recognition
• Word prediction
• Sentence comprehension forms units of

understanding

• Familiar context supports retention of read

material

Language and Numeracy

• Verbal codes for numerical concepts

– One; two; three….

• Word problems

– Tap existing language base

2 systems supporting learning = 2 possible deficits

Specific Working Memory Impairment (WMI) Specific Language Impairment (LI)

Strong Weak Strong Weak

Primary Deficits?

Mixed Impairments

Question 1

• Do specific and mixed deficits in language or

working memory occur in children?

Idea

• Examine language and WM skills of a large,

unselected group of young, school-age children.

Datasets

Archibald et al. (2013)

• 34 schools
• 1387 screened

– 85% English; 82% mother with some college education

• 392 assessments

– Language – Working memory – Nonverbal intelligence

Archibald & Joanisse (2009)

• 9 schools
• 400 screened

– 94% English

• 88 assessments

– Language – Working memory – Nonverbal intelligence

Definitions

• Language Impairment

< 86 on language composite

• Working Memory Impairment

< 86 on verbal & VSSP WM composite

• Possible profiles:

– SLI – SWMI – Mixed WM & LI – No deficits

CELF-IV (Semel et al., 2003); AWMA (Alloway, 2007)

Classification Results

Study 2 Study 1

Based on Archibald & Joanisse, 2009

Summary: Specific Impairments

• Children with

– SLI – SWMI – Mixed language and working memory deficits

• may,

– have different characteristics – respond differently to treatment – respond to different types of treatment

Question 2

• Does treatment aimed at language or working

memory result in domain-specific and/or cross-domain effects?

Investigating Treatment Effects

• Group design

– Clinical trial

Measuring Treatment Effects

• Treatment effects may be small effects in

impaired groups

• Detecting small effects depends on power

– The probability that a test will detect a difference if it’s there

• Power is low if sample size is low
• Sample size is often low

Single Subject Design

• Subject serves as his/her own control
• Evaluate the effect of intervention on

particular individual

• Uses same techniques as other clinical trials

– Blinding – Measures of control & target behaviours – Repetition across multiple individuals

SSD: Inferring causality

• Concomitant variation

– temporal arrangement

• baseline (control data)

– Determines expected level of performance

• Intervention (treatment data)

– Evaluate change in performance relative to expected

– copresence of intervention & change

• maintained over time

SSD: Inferring causality

10 20 30 40 50 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Baseline (Control data) Treatment

Sets the expected level

• f performance.

Defines the desired zone

SSD: Inferring causality

10 20 30 40 50 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Baseline (Control data) Treatment

Desired Zone 1/8 in desired zone 5/8 in desired zone

Proportions compared in a binomial distribution to determine if probability of intervention outcome is p<.05.

Co-presence of intervention & change

Idea

• Examine how children with language and/or

working memory impairment respond to language-focused or working memory-focused intervention using a SS design

Laura Pauls

Participants

• Recruited from existing database

– Based on language & WM testing at 2 time points (1 year apart); typical PIQ (or 10 point discrepancy) – Learning deficits; parent or teacher concern – 9-11 years old

• Profiles

– Cogmed: 5 SWMI; 2 L&WMI – Language: 8 SLI; 2 L&WMI

Intervention

Language-Focused

• Narrative based
• Focused on…

– Story structure – Story retelling – Vocabulary development – Grammatical complexity (based on individual abilities)

• Dosage

– 3x/wk for 5 wks – 40 minute sessions

Working Memory

• Cogmed
• Dosage

– 5x/wk for 5 wks – 30-40 minute sessions

• More comprehensive

Davies et al., 2004; Swanson et al., 2005; Westerveld & Gillon, 2008; Klingberg et al., 2005; Holmes et al., 2009

Study Design

Probes 2x/wk

Language or WM Intervention 3-5x/wk

Probes 1x/mo Week

4 9 13 26

Assessment

Probe Measures

Sentence Combining

“Selena flies her kite. It is not very windy.” “Selena flies her kite even though it is not very windy.”

Puzzle Completion Number Comparison Nonword Repetition

“da-moy-cho, tay-chee-dow, tow-doy-foo, voo-ta-yee” “tay-chee-dow”

Study Design

Probes 2x/wk

Language or WM Intervention 3-5x/wk

Probes 1x/mo Week

4 9 13 26

Assessment

Assessment Battery

• Working Memory

– AWMA (Alloway, 2007)

• Digit Recall, Counting Recall, Spatial Span
• Language

– CELF-IV (Semel et al., 2003)

• Recalling Sentences, Concepts & Following Directions
• Reading & Math

– TOWRE (Torgesen et al., 1999)

• Nonword reading, sight word reading

– WJ III

• Reading Fluency, Calculations, Math Fluency

Results

• Probes

– Sentence combining – Puzzle completion – Nonword repetition – Number comparison

• Standardized tests

Results: Sentence Combining

SWMI-1 SWMI-2 * SWMI-3 SWMI-5 LWMI-1 SWMI-4 LWMI-2

*

SLI-1 SLI-2 SLI-3

*

SLI-4 SLI-5 SLI-6 SLI-7 SLI-8 LWMI-3 LWMI- 4

* Propositional density Words and/or propositions per sentence

COGMED LANGUAGE

Results: Puzzle Completion

SWMI-1

* * *

SWMI-2 SWMI-3

* *

SWMI-4

*

SWMI-5

*

LWMI-1

* *

LWMI- 2 SLI-1 * SLI-2 SLI-3 SLI-4

*

SLI-5

*

SLI-6

*

SLI-7 SLI-8 LWMI-3 LWMI-4

*

# correct pieces chosen/second

COGMED LANGUAGE

Results: Nonword Repetition

SWMI-1 * SWMI-2 SWMI-3 SWMI-4

*

SWMI-5 LWMI-1 LWMI-2

* *

SLI-1 *

*

SLI-2 SLI-3 SLI-4 SLI-5 SLI-6 SLI-7 SLI-8 LWMI-3 LWMI-4

% syllables correct % phonemes correct

*

COGMED LANGUAGE

Results: Number Comparison

SWMI-1 SWMI-2 SWMI-3 SWMI-4 SWMI-5 LWMI-1 LWMI-2 SLI-1 SLI-3 SLI-4 SLI-5 SLI-6 SLI-7 SLI-8 LWMI-3 LWMI-4

% items correct

SLI-2

COGMED LANGUAGE

Probe Results: Cogmed

WM Task: Puzzle Completion Language Task: Sentence Combining Both: Select Nonword Repetition Control: Number SWMI-1 * * * SWMI-2 * * SWMI-3 * SWMI-4 * * SWMI-5 * * LWMI-1 * LWMI-2 * *

*significant increase relative to baseline during either intervention or follow-up

All participants improved Clear near-transfer effect; mixed cross-domain effects

Probe Results: Language Tx

WM Task: Puzzle Completion Language Task: Sentence Combining Both: Select Nonword Repetition Control: Number SLI-1 * * SLI-2 SLI-3 * SLI-4 * * SLI-5 * * SLI-6 * SLI-7 SLI-8 * * LWMI-3 * * LWMI-4

7/10 participants improved Mixed results – same & cross-domain effects

Results

• Probes

– Cogmed: near transfer; some cross-domain effects – Language: mixed effects – same & cross-domain

• Standardized tests

Standardized Test Results: Cogmed

Working Memory (AWMA) Language (CELF-IV) Reading (TOWRE / WJ- III) Math Fluency (WJ-III) SWMI-1 * SWMI-2 * SWMI-3 * * * SWMI-4 * SWMI-5 * * * LWMI-1 LWMI-2 * *

*>10 point increase on at least one subtest relative to baseline during either intervention or follow-up

Standardized Test Results: Cogmed

Working Memory (AWMA) Language (CELF-IV) Reading (TOWRE / WJ- III) Math Fluency (WJ-III) SWMI-1 ** SWMI-2 * * SWMI-3 ** * * SWMI-4 ** SWMI-5 ** * * LWMI-1 * LWMI-2 * * + Puzzle Probe

Standardized Test Results: Cogmed

Working Memory (AWMA) Language (CELF-IV) Reading (TOWRE / WJ- III) Math Fluency (WJ-III) SWMI-1 ** * SWMI-2 * ** SWMI-3 ** * * SWMI-4 ** SWMI-5 ** ** * LWMI-1 * LWMI-2 * * * + Puzzle Probe + Sentence Comp Probe

Same-domain effects; common cross-domain effects; limited far transfer to reading or math

Standardized Test Results: Language

Working Memory (AWMA) Language (CELF-IV) Reading (TOWRE / WJ-III) Math Fluency (WJ-III) SLI-1 * * * SLI-2 SLI-3 * * * SLI-4 * * * SLI-5 * SLI-6 * * SLI-7 * * SLI-8 * LWMI-3 * LWMI-4 * * Increase of 10 points in standard score, either post treatment or at follow-up

Standardized Test Results: Language

Working Memory (AWMA) Language (CELF-IV) Reading (TOWRE / WJ-III) Math Fluency (WJ-III) SLI-1 * * * SLI-2 SLI-3 * * * SLI-4 * * * SLI-5 * * * SLI-6 * * SLI-7 * * SLI-8 * * LWMI-3 * * LWMI-4 * + Puzzle + SC

Common same & cross-domain effects; some far transfer to reading

Results: Summary

• Cogmed & language-focused intervention

– Evidence for same-domain & cross-domain improvements – Transfer effects to reading more common after language intervention – Minimal transfer to math

4 Take-home Points

• 1. WM & long term language knowledge

support learning

• 2. Separable deficits in these systems occur

4 Take-home Points

• 3. Treatment targeting individual’s specific

deficits can lead to specific and generalized improvements

• 4. Cogmed

– Improved WM for all participants with WMI – Improved language performance for several

Thank you!

Research Students:

• Laura Pauls
• Areej Balilah
• Nicolette Noonan
• Monica DaSilva

LWM Lab

Collaborators:

• Marc Joanisse
• Daniel Ansari
• Janis Oram Cardy

Funding:

• Natural Sciences and Engineering

Research Council

• Ministry of Research & Innovation

Early Researcher Award

Questions?

• To contact me:

–larchiba@uwo.ca –Lab website

• http://www.uwo.ca/fhs/lwm/
• Follow our lab blog

– http://www.canadianslp.blogspot.com/

LWM Lab