Making the links Dr Anna Smith No task

Making the links Dr Anna Smith

No task is ‘pure’ • Neuropsychological functions operate together rather than in isolation • Theoretically, a pure task would lack ecological validity • We need to examine evidence • Certainty is not guaranteed

No single diagnostic test • Dyslexia is heterogenous • No different from other developmental disorders • eg. , ADHD is diverse with several: § Etiological pathways § Clinical presentations § Levels of severity

Two cases of ADHD Case 1: Dan (12) • Born prematurely • low scores on target detection tasks • Academically underachieving • Easy to manage Case 2: Sian (41) • Talks continuously and changes subject • Difficulty keeping job for very long • Several relationships in succession • Impulsive hobbies that don’t last long • Performance is average on most tasks

Dyslexia is also heterogenous • Dyslexia can be heterogenous • We must make judgement on the neuropsychological evidence we have gathered § Test scores § Background information § Qualitative information during assessment • A good assessor will make the links

Reading • Comparison of reading single words and reading in context § Higher scores in tests of comprehension suggest inference when reading in context • Comparison of timed and untimed reading of single words with WRAT and TOWRE 2 § poor reading fluency is more likely to be observed in older children and adults • Comparison with PA subtests § Errors on phonetically plausible words suggest phonological deficits

Spelling • Good readers can be poor spellers § They tend to make ‘phonetic’ errors eg. , spelling cough, as coff § Poor readers and poor spellers make more ‘non-phonetic’ errors eg, spelling cough as coft • Good spelling in DASH tests cannot rule out poor spelling § may be attributable to word selection § Link standard spelling test with spelling in DASH tests • Sound based errors are associated with phonological errors § More often observed in poor readers • Rule based errors are associated with word structure difficulties § More often observed in good readers

Writing • Fine motor problems will be seen in DASH tests § Note handwriting § Link with scores from developmental coordination disorder screening tools § Link with performance on WRIT diamonds but may be specific to handwriting § Compare copy best and copy fast § Free writing may deteriorate over time (qualitative observations important here)

Writing • Language difficulties will be observed in free writing § Compare scores with WRIT verbal subscores § May be consistently poor rather than decline over time • Don’t forget to look at spelling and word choice

Phonological awareness • Subtests may not be equally sensitive: § Elision subtest – involves repetition and allows for other visual strategies § Blending and phoneme isolation are purer measures of PA § Use of supplementary tests containing non-words • eliminates the influence of vocabulary thus a purer test • May be useful in adults • Acceptable for APC as long as case made

Phonological awareness • Scores may not be below average but… § Qualitative analysis important as tests are not timed § Which of the items are difficult: • Compound words: see[saw] - see • Syllables: ri[der] - rye • Onset and rime: ma[n] - ma • Phonemes: tan[k]s - tangs

Phonological awareness § Links can be made with TOWRE 2 (sight word subtest may be better than phonemic decoding subtest) § Links can be made with spelling • more on this later § Links with reading single words • Errors with phonetically plausible real words suggest phonological awareness problem

Phonological memory • Non-word repetition task (CTOPP 2) § CTOPP 2 version does not include many established word patterns § task is purer § Can be indicator of specific language impairment (Bishop 2006) • Link with forwards letters and digits (correlated strongly (Wijsman et al, 2000) § These are established, familiar sounds § If deficit here as well, then more generalised problem with short term memory

Phonological memory • What other clues might you be looking for? § Teacher reports § GORT comprehension subtest – low score may mean less able to maintain information § Specific to phonological information? Compare with manual imitation (TOMAL 2) – visual information may be better held

Working memory • TOMAL 2 is very useful as we can compare directly backwards and forwards scores § Manipulation of letters and numbers may be challenging § Individuals with dyslexia make reversals of number pairs and inversions (6 for 9) • Make links with PA sub-tests: elision requires word to be held and manipulated in WM • Make links with memory for stories (TOMAL 2) § Good for semantic memory

Speed of processing • Typically measured by Rapid Automatic Naming • Not seen in all dyslexic individuals • Possible subtype? • Speed and quality of visual information is compromised • Letter identification is slowed • Links between frequently occurring letters are not made

Speed of processing • May be related to orthographic skill rather than phonemic encoding • May mean slow access to phonological representations rather than visual input (cancellation task) or phonological output (articulation rate) • Regular and irregular words are equally challenging • Further reading: Wolf et al (2000) and Araujo et al (2010), Georgiou et al, 2013 • Links can be made with: § TOWRE: a good measure of reading fluency § WIAT reading speed § If generally slow processing - SDMT – speed of processing

Language difficulties • Background may reveal information: § parental report describes word finding difficulties ‘she knows the answer but she can’t find the word’ § Significant delay in acquisition of first words • Verbal subtest of WRIT may reflect account § requires a qualitative analysis – definitions may involve pointing and using a basic explanation • Analogies subtest of WRIT § a feel for the answer but a difficulty expressing it • Link with low score on phonological memory tests (Bishop 2006)

Language difficulties • Be aware of discrepancy between performance and verbal ability eg. , average versus 70 (Georgopoulos et al, 2003) § strict criteria but useful • Comprehension is often better than production • Qualitative information can be accessed through responses on WIAT (an advantage over WRAT) § Syntax is often impaired § Mean length of utterance is short

Attentional difficulties • Parent and teacher report is crucial here § Inability to concentrate, finish anything, disrupts games • Qualitative observations during testing also important • Use a screening tool to gather information • A great website is http: //www. sdqinfo. org/ • May not be clear indications on tests § Cognitive Variability is hallmark of ADHD § poor motivation, distractability during tests

Making links means comparing scores • Where possible use confidence intervals of 95% but be cautious with these • If you have subtests which differ you can use confidence intervals to be objective • Convert scaled scores to standard or quotient scores • Example: Comparison of digit forwards and manual imitation: § Digit forwards 7 = 85 (SEM is 3) 95% CI is 79 -91 § Manual imitation 10 = 100 (SEM is 3) 95% CI is 94 -106 § No overlap thus confidence performance is markedly discrepant § Avoid use of the word ‘significant’ as you are not performing a statistical test

Single subtests versus composites/ indexes • It can be misleading to focus upon one subtest § Especially if that subtest is not reliable (less than. 9 reliability) • APC preference is for composite scores where appropriate • BUT if subtests within a composite are discrepant, it makes no sense to ‘pool’ scores either § Use outcomes cautiously § Consider reliability of each test § Use qualitative data • If you don’t feel confident, just report composite scores CIs

Table of SEMS for CTOPP-2 and TOMAL 2

Using your own CIs • You may wish to use your own calculations but worry that they differ from those provided in the table • Use footnotes to make it clear *Confidence intervals were calculated using SEMs provided in manual * Confidence intervals were calculated using reliability coefficients provided in the manual

Suggested statements The scores for Test X and Test Y result in the lack of an overlap of the associated confidence intervals of 95%. This discrepancy suggests that the composite score associated with these tests has less meaning than the individual test scores themselves. There is a marked discrepancy between Test A and Test X as defined by the lack of an overlap of the associated confidence intervals of 95%. This suggests that Learner has difficulties with skill A in comparison with skill X It is best to avoid the word ‘significant’ as this implies a formal test of score differences which has not been carried out for these tests.