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
Recommend
More recommend
