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

Specific Learning Disabilities: The Role of Working Memory and Other Domain-specific Deficits Lisa Archibald, PhD Western University, Canada Cogmed Conference 2014

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

Specific Learning Disabilities Dyslexia Specific Language Impairment Dyscalculia

Specific Learning Disabilities • Heterogeneity – Different underlying cognitive impairments? • Domain-general – Memory, executive functioning, etc. • Domain-specific – Phonological processing, magnitude processing

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

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

Primary Deficits? Specific Working Specific Language Memory Impairment Impairment (WMI) (LI) Strong Strong Weak Weak 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) Archibald & Joanisse (2009) • 34 schools • 9 schools • 1387 screened • 400 screened – 85% English; 82% mother – 94% English with some college education • 88 assessments • 392 assessments – Language – Language – Working memory – Working memory – Nonverbal intelligence – 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 1 Study 2 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 60 Defines the 50 desired zone 40 Sets the expected level 30 of performance. 20 10 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Baseline (Control data) Treatment

SSD: Inferring causality Co-presence of 60 intervention & change 50 Desired Zone 40 30 20 10 1/8 in desired zone 5/8 in desired zone 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Baseline (Control data) Treatment Proportions compared in a binomial distribution to determine if probability of intervention outcome is p <.05.

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 Working Memory • Narrative based • Cogmed • Focused on… • Dosage – Story structure – 5x/wk for 5 wks – Story retelling – 30-40 minute sessions – Vocabulary development – Grammatical complexity • More comprehensive (based on individual abilities) • Dosage – 3x/wk for 5 wks – 40 minute sessions Davies et al., 2004; Swanson et al., 2005; Westerveld & Gillon, 2008; Klingberg et al., 2005; Holmes et al., 2009

Study Design Language or WM Assessment Intervention 3-5x/wk Probes 2x/wk Probes 1x/mo 26 Week 13 4 9

Probe Measures Puzzle Completion Nonword Repetition “da -moy-cho, tay-chee-dow , tow-doy-foo, voo-ta- yee” “ tay-chee-dow ” Sentence Combining Number Comparison “Selena flies her kite. It is not very windy.” “Selena flies her kite even though it is not very windy.”

Study Design Language or WM Assessment Intervention 3-5x/wk Probes 2x/wk Probes 1x/mo 26 Week 13 4 9

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

COGMED LANGUAGE Results: Sentence Combining SWMI-2 * SWMI-1 SLI-1 SLI-2 SLI-3 SLI-4 SWMI-3 SWMI-4 * SLI-5 SLI-6 SWMI-5 SLI-7 SLI-8 LWMI-1 LWMI-2 LWMI-3 LWMI- * 4 * Propositional density Words and/or propositions per sentence

COGMED LANGUAGE Results: Puzzle Completion SWMI-1 SWMI-2 SLI-1 * SLI-2 * * * SLI-4 SLI-3 SWMI-3 SWMI-4 * * * * SLI-5 SLI-6 * * SWMI-5 * SLI-8 SLI-7 LWMI-1 LWMI- 2 LWMI-3 LWMI-4 * * * # correct pieces chosen/second

COGMED LANGUAGE Results: Nonword Repetition SWMI-2 SWMI-1 * SLI-1 * SLI-2 * SLI-3 SLI-4 SWMI-3 SWMI-4 * SLI-5 SLI-6 SWMI-5 SLI-7 SLI-8 LWMI-2 LWMI-1 LWMI-4 LWMI-3 * * % syllables correct * % phonemes correct

COGMED LANGUAGE Results: Number Comparison SWMI-2 SWMI-1 SLI-1 SLI-2 SLI-3 SLI-4 SWMI-3 SWMI-4 SLI-5 SLI-6 SWMI-5 SLI-7 SLI-8 LWMI-2 LWMI-1 LWMI-4 LWMI-3 % items correct

Probe Results: Cogmed WM Task: Language Task: Both: Control: Puzzle Sentence Select Nonword Number Completion Combining Repetition 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: Language Task: Both: Control: Puzzle Sentence Select Nonword Number Completion Combining Repetition 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