A systematic review and meta-analysis of dichotic deficits in individuals with learning disability

ABSTRACT

Classic dichotic listening test using speech stimuli results in right ear advantage, due to the dominant crossed pathway for speech and language. However, such a conclusion cannot be derived for children with learning disability (LD) because of these variabilities among studies. Hence the study aimed at integrating the findings from various studies and synthesize evidence for the association between dichotic listening and hemispheric asymmetry in individuals with LD. All necessary key terms for ‘Learning disability’ and ‘dichotic listening’ were identified and used in combination with appropriate Booleans to obtain articles of interest. The initial search resulted in 3643 articles out of which 50 studies were selected for data extraction. The Right Ear Advantage (REA) is preserved in individuals with LD for the dichotic free recall test. However, in this condition, the right ear scores were significantly poor compared to controls. In the directed attention condition, the individuals with LD showed difficulty in right attended conditions whereas no deficit is observed in left attended conditions. Overall evidences indicated that individuals with LD showed poorer performance in both free and forced recall irrespective of dichotic material used indicating both top-down and bottom-up processing deficits.

KEYWORDS:

• Dichotic listening
• learning disability
• hemispheric asymmetry
• ear advantage

1. Introduction

One of the auditory processes that utilizes binaural hearing is dichotic listening (DL). In the classical dichotic listening task (Free recall condition), the listeners hear two distinct stimuli in each ear and are asked to recall these stimuli in any order. The performance of two ears typically differs due to binaural competition during this process (Hugdahl et al., 2009). According to Kimura's structural model (Kimura, 1967), listeners perceive the signal with the right ear better because of the Right Ear Advantage (REA), phenomena of underlying left hemisphere dominance in language processing. The signal in the right ear directly reaches the left hemisphere through the contralateral channel and is therefore perceived better (Blumstein, Goodglass, & Tartter, 1975).

It is commonly used to measure dichotic deficits and laterality in both experimental and clinical research. Since its inception, the dichotic test has undergone a number of revisions that assess higher cognitive processes including attention and executive function in addition to auditory processing. One such modification is the ‘forced recall’ test, in which participants are told to concentrate on either the right or left ear. Right ear advantage is typically the outcome of selectively focusing to the right ear (Forced Right (FR) state), and the opposite is true for the forced left (FL) condition. The cognitive and attention networks involved in the processing of auditory language are stated to be evaluated by the forced attention or directed attention dichotic circumstances of the DL test (Hugdahl & Andersson, 1986).

Learning disabilities (LD) are a diverse collection of disorders that are characterized by severe challenges in the development and use of cognitive skills such as speaking, listening, reading, writing, thinking, and math (Brunswick & Rippon, 1994). These conditions are believed to be caused by central nervous system dysfunction and are inherent to the individual (Gartland & Strosnider, 2004). These children exhibit a variety of linguistic and academic deficiencies (de Wit et al., 2016). Dyslexia is no longer an official term used under federal education law. This term was initially used to describe a language-based (i.e., phonologically-based) reading disability (Anastasiou & Polychronopoulou, 2009). Lack of linguistic lateralization utilizing both free recall and forced recall is revealed by DL testing in children with LD (Obrzut & Mahoney, 2011). Several research has used traditional and modified dichotic listening procedures to examine dichotic deficits in children with LD (Obrzut & Mahoney, 2011).

Most articles have noted a substantial REA or higher RE (Right ear) scores in children with LD as compared to the control group (Hakvoort et al., 2016; Hynd, Obrzut, Weed, & Hynd, 1979; Kershner, Henninger, & Cooke, 1984; Obrzut, Bryden, Lange, & Bulman-Fleming, 2001). However higher performance from LE (Left Ear) was also mentioned (Bowen & Hynd, 1988; Brunswick & Rippon, 1994; Helland & Asbjørnsen, 2001; Hugdahl, 2003) and Left Ear Advantage (LEA – Higher scores in left ear when compared to right) (Gupta, 2015; Smith & Griffiths, 1987). These patterns of performance change can vary depending on the type of dyslexia and the tasks (Aylward, 1984; Kołtuska & Grabowska, 1992; Patel & Licht, 2000). It is challenging to draw conclusions from the several studies that have used the DL test on children with LD because of the wide variations in the materials (stimuli material, scoring pattern, amount of items) and study outcomes. There are no systematic reviews or meta-analyses that objectively assess each study's ability to pinpoint the dichotic deficiencies in children with LD as of yet. To further examine the quality of the evidence and to ascertain the effect size, a meta-analysis is required. Therefore, the current review aims to consolidate data regarding the relationship between dichotic listening and hemisphere asymmetry in people with LD and combine findings from multiple investigations.

2. Method

This systematic review and meta-analysis were conducted in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). The current systematic review protocol was prospectively registered and approved by PROSPERO database (Register number: CRD42021228691).

2.1. Keyword build

Using the Cochrane library, Medical term (MeSH) search engine and also taking inputs from experts, all necessary key terms for ‘Learning disability’ and ‘dichotic listening’ were identified and used in combination with appropriate Booleans to obtain articles of interest. We have included dyslexia as a key term because we wanted to extract articles even before 1980 during which the term dyslexia was commonly used. The search string was as follows.

(((((Dichotic listening) OR (divided attention)) OR (directed attention)) OR (forced recall)) OR (free recall)) AND ((((((((((((((((((learning disability) OR (learning disabilities)) OR (learning disturbance)) OR (learning disturbances)) OR (learning disorder)) OR (learning disorders)) OR (developmental academic disability)) OR (developmental academic disabilities)) OR (developmental disorders of scholastic skills)) OR (developmental disabilities of scholastic skills)) OR (developmental academic disorder)) OR (developmental academic disorders)) OR (scholastic skills development disorder)) OR (scholastic skills development disorders)) OR (dyslexia)) OR (dyscalculia)) OR (specific learning disorder)) OR (specific learning disorders))

2.2. Search strategy

Twenty-three keywords were used to search five major databases including PUBMED, PROQUEST, SCOPUS, WEB OF SCIENCE & CINAHL. The search was predominantly run for the title and/or abstract for all articles published till 2020. 3643 articles were retained after removing duplicates (n = 1910) in Rayyan software.

2.3. Screening/selection process

A total of 5553 articles obtained from the online search engine were screened by the first (GG) and second (PP) authors for inclusion based on the selection criteria. The details of the screening are provided in the PRISMA flow chart. Any conflict in the decision between the first and the second authors was resolved by the third author (MB). Title and abstract screening, full-text screening, and Data extraction were done using Cochrane-recommended online systematic review software named ‘Ryaan’ (Khabsa, Elmagarmid, Ilyas, Hammady, & Ouzzani, 2016).

Articles were included for systematic review and meta-analysis if they [1] Cross-sectional studies, Case–control studies, and Cohort studies to date [2] Articles published in English [3] Presence of one or more mentioned keywords in the title or abstract. Articles were not included for systematic review and meta-analysis if they [1] Study designs other than case–control and cohort study [2] Use of quasi dichotic listening paradigm [3] Studies that are not published in peer-reviewed scientific journals [4] Systematic reviews within the topics were eliminated. Based on full-text screening by the reviewers, 50 articles were included for the narrative review.

2.4. Data extraction

Data extraction was carried out in a Microsoft word document by reviewer 1, GG. The extracted data included the following outcomes. The major outcome considered was the right and left ear scores for the mentioned dichotic test, the conditions at which the dichotic test is done.

The minor outcomes included type of study, study location, the population sample size for the control group (Children with out LD) and experimental group (Children with LD), mean age of the different groups, type of sampling done, measures of handedness, hearing, LD, and other tests included, including dichotic test/ stimuli used, transducers and its output level, recording person and equipment used to present, mean scores of the dichotic tests, response, major findings, and its justification.

2.5. Study quality assessment

The case–control version of the Newcastle Ottawa scale (NOS) was used for the evaluation of risk factor studies (Peterson, Welch, Losos, & Tugwell, 2011). The tool evaluates studies’ quality in terms of selection, exposure, and comparability of a total of eight items. Quality assessment was performed by three expert reviewers who are experienced in the field of electrophysiology (MB and HP). Two reviewers performed the quality assessment of the included studies. If the first two reviewers did not reach a consensus, the score of the third reviewer was decisive.

2.6. Meta-analysis

Those articles with necessary quantitative data were synthesized into a meta-analysis using Review Manager (version 5.2). Two meta-analyses were performed to find the dichotic listening deficits in children with LD. Firstly, the pooled mean right and left ear scores of both control subjects and children with LD from various studies are compared descriptively. Traditional meta-analysis was not possible in this case because of the lack of any standard formula or software to calculate the effect size within-subject variables. Second, a random-effects meta-analysis was done using the Mean Difference (MD) for the dichotic scores between the LD and control group with a 95% confidence interval. The results of the meta-analysis were evaluated based on the pooled evidence to calculate the overall effect (p-value).

3. Results

The PRISMA flowchart in Figure 1 illustrates the overall studies from the search and the different steps involved in excluding the irrelevant studies and finalizing the most related studies. The initial search in the five databases using the key terms excluding the detected duplicates resulted in 3643 articles. Title and abstract screening were done based on the inclusion and exclusion criteria and 3474 articles were removed. Hence, 169 articles proceeded for full-text screening, and 50 studies were selected for data extraction (Table 1).

Figure 1. Prisma flow chart explaining stages of meta-analysis.

Table 1. Summary of evidence from the systematic review.

Sl. No	Title/Author/Year	Participants characteristic (groups, sample size and mean age)	Dichotic test used	Transducer used and stimulus intensity	Responses	Conditions	Findings
1.	Hugdhal (2003)	CG-23 (20B & 3G) Dys G-23 (19B & 4G) MA; - Age range; 10–12 yrs	DCV 6 CV syllables /ba/, /da/, /ga/, /ka/, /pa/, /ta/ Total stimuli; 108 Controlled duration: yes	Intensity – 75 dB SPL	Verbal	FR and DA	Reduced right ear scores in Dys group
2.	Kołtuska, B & Grabowska, A (1992)	CG: 12 (7B & 5G) Dys G: 29 -phonetic disorder: 17 (10B & 7G) Language disorder: 11 (6B & 5G) MA: na Age range: 8–9 years	Exp 1: Two pairs of words separated by 500 ms interval (108 pairs) Controlled duration: NA Exp 2: One pair of two different words (114 pairs)	Conversational level lower than conversational level	Verbal	FR	Significantly higher scores for right ear in all three groups Mean recognition level for both ears was lesser for dyslexics group
3.	Brunswick, N & Rippon, G (1994)	CG: 15 (B) MA: 8.78 ± 0.95 Dys G: 15 (B) MA: 8.87 ± 1.41	DCV Controlled duration: NA 32 dichotic pairs 320 ms duration	Intensity – 75 dBSPL	verbal	FR	Weaker right ear performance by dys group No difference in left ear performance between groups
4.	Caplan, B., & Kinsbourne, M. (1982)	Control group: 40 (21B & 19G) MA – 9.26 ± 1.41 Poor readers: 40 (36B & 4G) MA – 13.39 ± 2.24	DCV CV syllables /ba/,/da/,/ga/,/ka/,pa/,/ta/ ISI, 4msec	Intensity: 50 dBSPL	verbal	FR	No significant degree of dichotic ear asymmetry between the groups
5.	Dermody, P., Mackie, K., & Katsch, R. (1983)	CG: 15 MA – 11.5 years LD group: 15 MA – 12.7 years Gender Information: NA	DCV /pa, ta, ka, ba, da, ga/ 90 msec	Intensity: 80 dB SPL	verbal	FR	No right ear advantage in both groups Differences for the groups are observed only in double-correct responses.
6.	Hakvoort, B., de Bree, E., van der Leij, A., Maassen, B., van Setten, E., Maurits, N., & van Zuijen, T. L. (2016)	CG: 26 DD; FRND: 26 FRD: 17	Dichotic digit test Grade 3 (3 monosyllabic digits /ear) Dichotic listening test Grade 6 (/bɑ/, /dɑ/, /ɡɑ/, /pɑ/, /tɑ/ and /kɑ/)	70 dB 80dB Duration controlled: yes	Verbal	FR & DA	Right ear is dominant FRD group; lower left ear scores show lager REA
7.	Harel, S., & Nachson, I. (1997)	CG: 63 Poor readers: 60 MA: na Age range: 8.9 -12.9 yrs	Pairs of tonal stimulus sets (64 pairs) – three successive high- or low-frequency tones Duration: 600 msec Duration controlled: yes	Intensity: NA	Verbal	FR & DA	Correct responses were highest in free recall and directed attention along with presence of continuous tone in the unattended ear
8.	Helland, T., & Asbjørnsen, A. (2001)	CG: 20 (16B & 4G) MA: 12.11 ± 0.36 Dys G: 43 (36B & 7G) MA: 12.67 ± 1.68	CV syllables (108) Duration controlled: NA	Intensity: NA	Verbal	FR&DA	Weak left ear dominance in dys group and that of right ear in CG
9.	Asbjørnsen, A. E., & Helland, T. (2006)	CG: 20 MA: 12.11 ± 0.36 LD group: 43 (RI/NLI – 25 RI/LI – 18) MA: 12.67 ± 1.68	DCV CV syllables /ba/,/da/,/ga/,/ka/,pa/,/ta/ Duration controlled: no	Intensity: NA	Verbal	FR&DA	Reduced left hemisphere language dominance with reading impairment
10.	Kershner, J., & Micallef, J. (1991)	CG: 30 (boys) MA: 7.8 ± 0.9 LD group: 30 (boys) MA: 11.3 ± 1.1	DDT 350msec duration Duration controlled: NA	> 70dB	Verbal	DA	Equal number of subjects showed REA in both groups
11.	Helland, T., Asbjørnsen, A. E., Hushovd, A. E., & Hugdahl, K. (2008)	Control group: 40 (29B & 11G) MA: 12.40 ± 0.58 Dys group, D1: 20 (10B & 10G) MA: 12.86 ± 0.65 D2: 20 (14B & 6G) MA: 12.59 ± 0.92)	DCV syllables	Intensity: NA	Verbal	FR	-Significantly more correct responses from the right ear -REA was absent in the D2 group compared to D1 and CG
12.	Hugdahl, K., Heiervang, E., Nordby, H., Smievoll, A. I., Steinmetz, H., Stevenson, J., & Lund, A. (1998)	Control group: 25 (Mean age 11.7 ± 0.5) Dyslexic group: 25 (Mean age 11.8 ± 0.4)	CV syllables /ba/,/da/,/ga/,/ka/,pa/,/ta/ 36 dichotic pairs and 36 trials Duration: 400msec	(75 dB)	verbal	FR&DA	Significant REA in both groups during DA (R) and no significant for DA (L)
13.	Hynd, G. W., Obrzut, J. E., Weed, W., & Hynd, C. R. (1979)	CG: 48 (40B & 8G) LD group: 48 (40B & 8G) younger group of 24 children MA: 8.3 ± 8.19 24 older children MA: 10.6 ± 7.46	DCV 30 pairs Controlled duration: NO	55 dBHL	verbal	FR	Significant RE effect in both the groups
14.	Hynd, G. W., & Obrzut, J. E. (1981)	(3 groups of 30) CG: 15 LD group: 15 Younger group, MA: 8 ± 4.78 Intermediate group: 9.5 ± 6.56 Oder group: 11 ± 5.12	DCV 30 pairs- /ba/ ka/da/ga/ta/pa/ Controlled duration: NO	Intensity: NA	verbal	FR	Significant RE effect in both groups
15.	Kershner, J., Henninger, P., & Cooke, W. (1984)	Exp: 1 CG: 15 (B) Dys group: 15 (B) MA: na Age range: 10–12 yrs	84 dichotic pairs of monosyllabic digits (36 trails) single, double, triple, and quadruple pairs duration: 306msec Controlled duration: NA	70 ± 5 dBHL	Written and verbal	FR	Significant REA for CG Dys group produced a significant LEA The dys-phonetic children showed RH (hemisphere)A when responding in writing, condition
16.	Kershner, J. R., & Morton, L. L. (1990)	Exp:1 CG: 32 (27B & 5G) MA: 12.0 ± 1.0 LD group: 32 (27B & 5G) MA: 12.3 ± 1.4 Exp:2 Control group: 14 LD group:14	-64 dichotic pairs of monosyllable digits -two blocks (8 trials each) of quadruple pairs Duration: 307msec	65 dBHL	Written and Verbal	FR&DA	LD group-maintained REA in the right-first condition. (both exp)
17.	Kershner, J., & Micallef, J. (1991)	CG: 34 Dys group: 12 MA = 10.8 ± 1.2	350-msec duration 64 dichotic pairs of monosyllable digits duration controlled: NA	70 dBHL	Verbal	FR	Reading condition, greater REA in all three groups
18.	Kershner, J., & Micallef, J. (1992).	CG: 30 MA: 11.6 ± 1.1 DyS group: 30 MA: 11.3 ± 1.1	Dichotic CV Syllables quasi-random arrangement duration controlled: NA	70 dBHL	Verbal	FR&DA	Superior talents in LE recall which implicates a right hemisphere skill in dys group
19.	Keefe, B., & Swinney, D. (1979)	CG: 19 DD group: 19 MA: 10.0 yrs	Dichotic digits 2 digits pair per second Duration controlled: yes	Intensity – NA	Pointing	DA	RE superiority by both groups
20.	Kershner, J. R. (2016)	CG: 34 (6B & 28G) MA: 33.5 ± 9.3 Dys group: 73 Mod group: 42 (21B & 21G) MA: 26 ± 6.5 Sev group: 31 (19B & 12G) MA: 27.2 ± 7.9	Dichotic CV Syllables Duration controlled: yes	70 dBHL	Written	DA	LF, significantly greater REA by CG compared to mod group RF, lower REA by CG compared to sev group Poorer ability to shift from right to left
21.	Koomar, J. A., & Cermak, S. A. (1981)	CG: 30 (23B & 7G) MA: 9.0 yrs LD group: 30 (23B & 7G) MA: 8.10 yrs	DCV syllables DDT (20 series of 3 pairs of digits) Duration controlled: NA	74 dBSPL	Verbal	FR	Similar scores for LD group and CG on CV but significantly lower than the normal group on the digts
22.	Gupta, Pradip kumar (2015)	CG: 10 (9B & 1G) Dys group: 10 (7B & 3G) MA: na Age range: 10–12 yrs	DCV syllables Duration controlled: yes	40 dB SL	Verbal	FR & DA	Dys group: LEA observed Poorer performance on DCV test for single correct score of right ears, left ear and of double correct scores
23.	Martínez, J. A., & Sánchez, E. (1999)	CG: 75 (B) (age matched: 50 Reading level matched: 25) MA: 9.93 ± 0.72 Dys group: 50 (B) MA: 10.24 ± 0.84	DCV syllables Duration controlled: no	Intensity: NA	verbal	FR&DA	No significant differences were found for the FR and DA
24.	McKeever, W. F., & VanDeventer, A. D. (1975)	CG: 9 MA: 13.7 Dys group: 9 MA: 13.9	DDT 3 pairs	Intensity: NA		FR	LD possess left hemisphere language specialization and clear auditory memory deficits
25.	Menashe, S. (2019)	CG: 22 (11B & 11G) MA: 25.4 ± 3.35 Dys group: 18,(9B & 9G) MA: 26.4 ± 4.37	Dichotic CV syllables, Duration: 380ms Dichotic Vowels Task, 200 dichotic Vowels, 320ms	75dBHL	Pressing the button	DA	Differently organized auditory system with altered speech lateralization in adult dyslexic readers
26.	Moncrieff, D. W., & Black, J. R. (2008)	CG: 10 (7B&3G) MA: 11.8 ± 0.4 Dysc group: 10 (7B&3G) MA: 11.6 ± 0.3	DDT DCV Competing word subtest (50 pairs of single syllables words)	50 dB SL Duration controlled: NA	verbal	FR&DA	67%dys children showed clinically significant left ear weaknesses on DDT-FR Reduced right ear performance with CVs has been reported for children with dyslexia
27.	Morton, L. L. (1994)	CG: 20 (B) RD group: 40 (B) (4 subtypes) MA: na Age range: 9–12 years	Dichotic CV syllables, mixed order Duration controlled: no	70 dBHL	Verbal	FR&DA	RD group showed stronger REA on bilabials compared to alveolars and velars
28.	Milberg, V. W., Whitman, R. D., & Galpin, R. (1981)	CG:12 (B) MA: 10.0 yrs LD group: 12 (B) MA: 9.8years	Dichotic digit test -One digit of each pair Duration controlled: NA	Intensity: NA	Verbal	FR	Left ear yielded fewer errors than the right ear when it was the stored channel.
29.	Swanson, H. L. (1987)	CG: 16 (8B & 8G) MA-11.52 ± 0.94 LD group: 16 (8B & 8G) MA- 11.85 ± 0.69	Dichotic words (12 mono-syllabic or Bi-syllabic pairs) Duration Controlled: Yes	60 dB HTL	verbal	FR& cued recall	Recall was significantly higher for words in the LE than in the RE CG, immediately available for recall than LD group, irrespective of word feature
30.	Swanson, H. L., & Mullen, R. C. (1983)	CG: 24 (B) MA:11.63 ± 0.68 LD group: 24 (B) -older LD, MA: 12.2 ± 0.24 -younger LD, MA:8.75 ± 1.30	Nine words lists (3 semantic, 3 phonemic, 3 unrelated) in competing noise -12 monosyllable words Duration Controlled: Yes	60 dB HTL S/N ratio: 10	Verbal	FR&DA	LD group recalled more words on the both ear (left ear than right ear) LD group prefers semantically organized words in the LE, REA for phonemically organized and unrelated word lists for the older age.
31.	Pinheiro, F. H., Oliveira, A. M. D., Cardoso, A. C. V., & Capellini, S. A. (2010)	CG: 20 LD group: 20 MA: na Age range: 8–12 years	DDT (20 digits pairs, 4 digits to both ears) Portuguese-Language Alternate Dissyllable Dichotic test (SSW)	50 dB SL	verbal	FR&DA	Mean performance of LD group was lower than CG (both ears)
32.	Patel, T. K., & Licht, R. (2000)	CG: 20 (B) MA: 10.87 LD group; 33 (27B & 6G) P dyslexics: 14 MA: 10.38 yrs L dyslexics: 19 MA: 11.45 yrs	2 syllable Dutch words, four different affective tones, 144 pairs Duration controlled- yes	Normal speech loudness	Button press	FR	No significant ear advantage for P-types, and a REA for L-types and controls
33.	Obrzut, J. E., Bryden, M. P., Lange, P., & Bulman-Fleming, M. B. (2001)	CG: 72 (12B & 12G/grade) I, MA: 6.3 yrs III, MA: 8.2 yrs V, MA: 10.3 yrs LD group: 30 Age range: 8.0–11.10 yrs	2 syllable words, four different emotional tones, 144 stimulus pairs Duration controlled – yes	-	Verbal	FR	LD group, better RE scores than LE for words target
34.	Aylward, E. H. (1984)	CG: 34 LD group: 72 Dys-eidectics: 20 Dys-phonetics: 12 Non-specific dys: 20 MA: na Age range: 7–12 yrs	DDT 2 digits pair per second, 10 sec pause b/w trials	Constant volume	verbal	DA	Significant REA exhibited by dys-phonetics group, but not by CG and dys-eidetics group
35.	Beaumont, J. G., Thomson, M., & Rugg, M. (1981)	CG: 26 (B) MA (in months): 118.6 ± 10.2 LD group: 26 (B) MA: 118.6 ± 10.4	DDT Duration controlled: yes	Intensity: NA	Verbal	FR	Both groups showed a REA, and there was no overall difference between the groups in performance
36.	Obrzut, J. E., Obrzut, A., Bryden, M. P., & Bartels, S. G. (1985).	CG: 16 (14B & 2G) LD group: 16 (14B & 2G) MA: na Age range: 7.8–12.1 yrs	DCV syllables (30 Pairs)	55dBSPL	verbal	FR&DA	Greater dichotic recall for CG LD group, no consistent REA More correct responses from LE for DA (L)
37.	Bowen, S. M., & Hynd, G. W. (1988).	CG: 24 (17B & 7G) MA: 21.98 yrs LD group: 24 (17B & 7G) MA: 22.07 yrs	DCV syllables, Dichotic words, Dichotic sentences, Duration controlled: NA	Intensity: NA	Verbal	FR&DA	Both groups, REA present Most accurate responses for D words and DCV had least accurate scores DA(R) most correct responses
38.	Obrzut, J. E. (1979)	CG: II:19 IV: 27 Non-specific: II: 10 IV: 8 Dys group: (Dysphonetic Dyseidetic, Alexic) II: 43 IV: 37	DDT Duration controlled: yes	60 dB SPL	-	FR&DA	CG and dys-eidetic group showed better recall than dys-phonetic & alexic group significant REA only for dys-phonetic
39.	Obrzut, J. E., Conrad, P. F., Bryden, M. P., & Boliek, C. A. (1988).	CG: 15(10B & 5G) LD group: 15(8B &7G) LH group: 15(9B & 6G) Bilingual children: 15(8B &7G) MA: 9.9 yrs	DCV syllables, 3 sets and 30 trails Duration controlled: yes	55 dB SPL	Verbal	FR&DA	Constant REA by CG for all conditions LD group, REA for FR and DA(R) No ear differences for DA(L)
40.	Obrzut, J. E., Conrad, P. F., & Boliek, C. A. (1989)	CG:30(1B & 19G) LD group: 30(20B &10G) MA: 10.5 ± 1.8	DCV syllables non-verbal stimuli (serial presentation of two tones) – high/low Duration controlled: NA	70 dBSPL	Verbal	FR&DA	Right-handed reading-disabled group did not demonstrate a consistent REA More tonal stimuli recalled than CV, all groups More LE scores than RE
41.	Obrzut, J. E., Mondor, T. A., & Uecker, A. (1993)	CG: 34(15B &19G) MA: 9.2 ± 1.4 LD group: 17 MA:10.3 ± 1.1	DCV syllables (ba, da, pa, ta) Duration; 350 msec Duration controlled: NA	65 dBHL	Verbal	DA	Poor performance for DA(L)
42.	Morton, L. L., & Siegel, L. S. (1991)	CG: 30 LD group: 30 (28B & 2G) MA: 12.20 ± 1.39	DDT Duration, 306 msec (64 dichotic pairs of monosyllabic digits), 8 quadruple pairs	70 dBHL	Writing	FR&DA	No differences in laterality and overall recall performance LD & CG in the afternoon. Morning testing, CG showed greater REA for DA(R) and LEA for DA(L)
43.	Morton, L. L., & Kershner, J. R. (1993)	CG: 30 MA: 11.72 ± 0.74 LD group: 26(B) MA:11.24 ± 0.99	DDT 307 msec duration Eight quadruple Pairs	70 dB HL	Writing	FR&DA	Enhanced left hemisphere involvement after left hemisphere priming in the morning, and after right hemisphere priming in the afternoon
44.	Morton, L. L., & Siegel, L. S. (1991)	CG: 40 MA; 10.8 ± 1.1 LD group: 20 (comprehension) MA: 10.6 ± 0.9 LD group: 20 (comprehension and word recognition) MA:10.8 ± 0.7	DDT 307 msec duration 32 dichotic pairs (quadruple pairs) Duration controlled: NA	Intensity: NA	verbal	FR&DA	Reading disabled groups showed lower left ear advantage
45.	Boliek, C. A., Obrzut, J. E., & Shaw, D. (1988)	Control group: 25 (17B & 8G) MA:9.10 ± 1.7 LD group: 25 (17B & 8G) MA:10.1 ± 2.05	DCV Familiar CVC syllables & unfamiliar CV syllables	70-dB SPL	Verbal	FR&DA	Learning-disabled children did not demonstrate a consistent REA across all experimental conditions
46.	Yeni-Komshian, G. H., Isenberg, D., & Goldberg, H. (1975)	CG: 19 (9B & 10G) MA: 11.8 yrs LD group:19 (11B & 8G) MA: 12.8 yrs	DDT (8 digits except 6 and 7) 1/sec rate Duration controlled: yes	Normal conversational level	Verbal	DA	Overall, REA for both groups difference between the R and LE score was larger for LD group than CG
47.	Smith, K., & Griffiths, P. (1987)	CG: 12 (10B & 2G) MA: 11.36 yrs Dys group: 12 (10B & 2G) MA: 11.56 yrs	14 test pairs of non-phonological sounds.	65–75 dB	Verbal	FR&DA	-Significant LEA -Dyslexic children are unable to counteract the natural orienting tendency of their intact hemisphere
48.	Hugdahl, K., Helland, T., Færevaag, M. K., Lyssand, E. R. T., & Asbj rnsen, A. (1995)	CG: 53 Dys group: 53	DCV syllables 36 dichotic pairs	75 dB SPL	Pointing (selecting from a group)	FR	-Right-handed dyslexic subjects did not show the expected right ear advantage – left-handed group, there were no significant ear differences

CG – Control group, LD – learning disability group, RD – reading disability group, Dys – Individuals with dyslexia, B – Boys, G – Girls, MA – mean age, DCV – Dichotic CV, DDT – Dichotic digit test, FR – Free recall, DA – directed attention, REA – Right ear advantage, LEA – Left eat advantage, P – dyslexics; perceptual dyslexics, L – dyslexics; linguistic dyslexia, FRND – familial risk group of children who do not develop dyslexia, FRD – familial risk group of children who develop dyslexia, DD – Developmental dyslexia, RI/NLI – Reading impairment without impaired language comprehension, RI/LI – Reading impairment with impaired language comprehension, NA – not available.

3.1. General characteristics

3.1.1. Age and gender

The current review consisted of a total of 48 cross-sectional studies and 45 of these studies included participants of school age (8–13 years) except three studies which were on adults (23–32 years) (Bowen & Hynd, 1988; Kershner, 2016; Menashe, 2019). This review extracted data from a total population of 1348 age-matched controls and 1437 individuals with LD out of which 195 were adults (80 in control and 115 in experimental). The mean age for the control experimental group participants was 26.96 ± 6.32 and 25.41 ± 6.25, respectively. There were 630 male and 221 female participants in the control group and, 654 males and 168 females were included in the LD group. In two studies, control group was separated as age-matched controls (N = 70, MA = 10.7±1.1) and reading level matched controls (N = 45, MA = 8.7±2) (Martínez & Sánchez, 1999; Morton & Kershner, 1991). 18/48 studies included have not provided a gender-specific sample size (Hakvoort et al., 2016; Harel & Nachson, 1997; Helland & Asbjørnsen, 2001; Hugdahl et al., 1998; Hynd et al., 1979; Keefe & Swinney, 1979; J. Kershner & Micallef, 1991, 1992; Kershner & Morton, 1990; Obrzut et al., 2001; Obrzut, Hynd, Obrzut, & Pirozzolo, 1981; Pinheiro, Oliveira, Cardoso, & Capellini, 2010). The dichotic performance could vary as increasing age. Older dyslexic individuals have better scores when compared to their younger ones (Keefe & Swinney, 1979).

3.1.2. Hearing status

Routine pure tone audiometry and screening audiometric testing was used in 23 studies. Three studies have also used self/teacher reports to estimate the presence or any history of hearing loss (Aylward, 1984; Harel & Nachson, 1997; Obrzut et al., 1981). The hearing thresholds of the participants across the studies ranged between 15 and 20 dBHL for frequencies 250 Hz to 8000 Hz.

3.1.3. Handedness and IQ

Forty-six studies included right-handed individuals and two studies also included left-handed individuals (Boliek, Obrzut, & Shaw, 1988; Hugdahl, Helland, Færevaag, Lyssand, & Asbj Rnsen, 1995). A total number of 943 right-handed control and 105 left-handed children were included. Edinburgh Handedness Inventory (EHI) or its modifications were used to assess handedness (Oldfield, 1971 1971). Four studies have measured handedness using Annett's handedness inventory/questionnaire (Brunswick & Rippon, 1994; Hugdahl, 2003; Hugdahl et al., 1998; Moncrieff & Black, 2008). Few studies have also used Zazzo test and Aston Index (Hugdahl et al., 1995; Kołtuska & Grabowska, 1992) for the same. Some of the studies do not specify the tool used for assessment though they mention the handedness of the participants (Asbjørnsen & Helland, 2006; Bowen & Hynd, 1988; Hakvoort et al., 2016; Helland & Asbjørnsen, 2001; Kershner et al., 1984; Kershner & Morton, 1990; Koomar & Cermak, 1981; McKeever & VanDeventer, 1975; Menashe, 2019; Smith & Griffiths, 1987). Right-handed control groups showed an expected REA whereas the dyslexic group showed an equal performance on the right and left ear stimulus presentations. And for the left-handed individuals, no significant ear differences were reported (Hugdahl et al., 1995).

Similarly, 30 studies have quantified IQ using Wechsler Intelligence Scale for Children (WISC) or its revised version. Very few studies have used Peabody Picture Vocabulary Test (PPVT) alone and along with WISC, to quantify IQ (Keefe & Swinney, 1979; Obrzut et al., 2001; Obrzut, Conrad, & Boliek, 1989). Some studies have used different assessment tools such as Raven colored progressive matrices and Polish verbal intelligence, Otis Lennon mental age test, Slossen Intelligence test, and Quick test of Intelligence (Aylward, 1984; Kołtuska & Grabowska, 1992; McKeever & VanDeventer, 1975; Yeni-Komshian, Isenberg, & Goldberg, 1975). The structured interview has been used to screen the cognitive ability of the participants (Hugdahl et al., 1995). The average IQ considered was greater than or equal to 85 and not exceeding 115 with a mean IQ score of 94.95 for the experimental group and 111.33 for the control groups.

3.1.4. Stimuli used

In the current review, 45/48 studies have used verbal stimuli for dichotic assessment. The commonly used verbal stimuli are CVs, /ba/, /da/, /ga/, /ka/, /pa/, /ta/ and monosyllabic digits. 22 studies have used only dichotic CV testing (Asbjørnsen & Helland, 2006; Boliek et al., 1988; Brunswick & Rippon, 1994; Caplan & Kinsbourne, 1982; Gupta, 2015; Helland & Asbjørnsen, 2001; Hugdahl, 2003; Hugdahl et al., 1995; Hugdahl et al., 1998; Hynd et al., 1979; Kershner, 2016; Kershner & Micallef, 1992; Martínez & Sánchez, 1999; Menashe, 2019; Morton, 1994; Obrzut et al., 1981; Obrzut et al., 1989; Obrzut, Mondor, & Uecker, 1993; Obrzut, Obrzut, Bryden, & Bartels, 1985) whereas 13 studies used only DDT (Aylward, 1984; Beaumont, Thomson, & Rugg, 1981; Keefe & Swinney, 1979; Kershner et al., 1984; Kershner & Micallef, 1991; Kershner & Morton, 1990; Kershner & Stringer, 1991; McKeever & VanDeventer, 1975; Milberg, Whitman, & Galpin, 1981; Morton & Kershner, 1991; Obrzut, 1979; Yeni-Komshian et al., 1975). One study has assessed the dichotic performance using both the CV and digits (Koomar & Cermak, 1981). Apart from these, eight studies also have used sentences and words delivered in different emotional tones separately and along with CV syllables and digits (Bowen & Hynd, 1988; Kołtuska & Grabowska, 1992; Moncrieff & Black, 2008; Obrzut et al., 2001; Pinheiro et al., 2010; Swanson & Mullen, 1983). The remaining three studies have used either non-verbal/ tonal stimuli or the combination of CV/ digits with tonal stimuli (Harel & Nachson, 1997; Obrzut et al., 1989; Smith & Griffiths, 1987).

3.1.5. Transducer

All investigations, with the exception of one (insert earphones), employed headphones made by various manufacturers to deliver the dichotic stimuli. The majority have presented the stimuli above the threshold level or at the conversational level (mean intensity level – 66.86 dB SPL). 13 studies have not provided intensity-related information (Asbjørnsen & Helland, 2006; Beaumont et al., 1981; Bowen & Hynd, 1988; Harel & Nachson, 1997; Helland & Asbjørnsen, 2001; Keefe & Swinney, 1979; Martínez & Sánchez, 1999; Milberg et al., 1981; Morton & Kershner, 1991; Obrzut et al., 1981; Obrzut et al., 2001).

3.2. Outcomes

The current review focused on identifying the dichotic ear deficits in individuals with LD. Hence all the studies that used dichotic listening tests in the LD population have been included in the review. Forty-five studies have given the mean values of the correct responses in each ear for the dichotic tests being used while two studies also provided the double correct responses. Mainly three different dichotic paradigms were used across the studies. Fifteen studies tested only free recall condition (Beaumont et al., 1981; Brunswick & Rippon, 1994; Caplan & Kinsbourne, 1982; Hugdahl et al., 1995; Hynd et al., 1979; Kershner et al., 1984; Kershner & Micallef, 1991; Kołtuska & Grabowska, 1992; Koomar & Cermak, 1981; McKeever & VanDeventer, 1975; Milberg et al., 1981; Obrzut et al., 1981; Obrzut et al., 2001). Eight studied directed attention only (Aylward, 1984; Helland & Asbjørnsen, 2001; Keefe & Swinney, 1979; Kershner & Stringer, 1991; Menashe, 2019; Obrzut et al., 1993; Yeni-Komshian et al., 1975). The remaining 23 have included both free and forced recall on the same participants (Asbjørnsen & Helland, 2006; Boliek et al., 1988; Bowen & Hynd, 1988; Gupta, 2015; Hakvoort et al., 2016; Harel & Nachson, 1997; Hugdahl, 2003; Kershner & Micallef, 1992; Kershner & Morton, 1990; Martínez & Sánchez, 1999; L. Morton, 1994; Morton & Kershner, 1991, 1993; Obrzut, 1979; Obrzut et al., 1985; Obrzut et al., 1989; Obrzut, Conrad, Bryden, & Boliek, 1988; Pinheiro et al., 2010; Smith & Griffiths, 1987; Swanson & Mullen, 1983).

3.2.1. Free recall paradigm (FR)

3.2.1.1. Ear difference

Eighteen studies suggested the presence of right ear advantage in children with LD in the free recall paradigm (Beaumont et al., 1981; Dermody, Mackie, & Katsch, 1983; Hakvoort et al., 2016; Helland, Asbjørnsen, Hushovd, & Hugdahl, 2008; Hugdahl, 2003; Hugdahl et al., 1998; Hynd et al., 1979; J. Kershner, 2016; Kershner & Micallef, 1991, 1992; Kołtuska & Grabowska, 1992; Koomar & Cermak, 1981; Martínez & Sánchez, 1999; McKeever & VanDeventer, 1975; Obrzut, 1979; Obrzut et al., 1981; Obrzut et al., 1988; Obrzut et al., 2001). Ten studies found reverse ear asymmetry, ie LEA LDs (Brunswick & Rippon, 1994; Harel & Nachson, 1997; Hugdahl et al., 1995; Kershner et al., 1984; Milberg et al., 1981; Morton & Kershner, 1991; Morton & Siegel, 1991; Obrzut et al., 1989; Smith & Griffiths, 1987; Swanson & Mullen, 1983).

3.2.1.2. Performance difference

Eighteen studies suggest that children with LD significantly underperformed compared to controls irrespective of the ear (Beaumont et al., 1981; Brunswick & Rippon, 1994; Hakvoort et al., 2016; Harel & Nachson, 1997; Hugdahl, 2003; Hugdahl et al., 1998; Hynd et al., 1979; Kershner et al., 1984; Kershner & Micallef, 1991; Kołtuska & Grabowska, 1992; Martínez & Sánchez, 1999; McKeever & VanDeventer, 1975; Milberg et al., 1981; Moncrieff & Black, 2008; Obrzut & Mahoney, 2011; Smith & Griffiths, 1987; Swanson & Mullen, 1983). However, five studies suggested better performance in the experimental group (Bowen & Hynd, 1988; Dermody et al., 1983; Koomar & Cermak, 1981; Smith & Griffiths, 1987; Swanson & Mullen, 1983), this is commonly seen in dichotic tests with CVs and words compared to other stimuli (Bowen & Hynd, 1988; Koomar & Cermak, 1981). Also, no significant degree of ear or performance asymmetry between the groups was reported in three articles (Beaumont et al., 1981; Caplan & Kinsbourne, 1982; Kershner & Micallef, 1992).

3.2.2. Directed attention paradigm (DA)

3.2.2.1. Ear difference

Overall, studies have shown that attention improves scores in the attended ear. Twenty studies have suggested better RE scores attended right condition when compared to free recall right scores (Beaumont et al., 1981; Boliek et al., 1988; Bowen & Hynd, 1988; Hakvoort et al., 2016; Helland & Asbjørnsen, 2001; Hugdahl et al., 1998; Keefe & Swinney, 1979; Kershner & Morton, 1990; Kershner & Stringer, 1991; Martínez & Sánchez, 1999; Menashe, 2019; Morton, 1994; Morton & Kershner, 1993; Morton & Siegel, 1991; Obrzut, 1979; Obrzut et al., 1985; Obrzut et al., 1989; Obrzut et al., 1993; Yeni-Komshian et al., 1975), except one study (Dermody et al., 1983). For attended left condition, 10 studies have found greater responses in LE (Bowen & Hynd, 1988; Kershner & Micallef, 1992; Kershner & Morton, 1990; Kershner & Stringer, 1991; Morton & Kershner, 1991; Morton & Siegel, 1991; Obrzut et al., 1985; Obrzut et al., 1989; Smith & Griffiths, 1987; Swanson & Mullen, 1983) whereas six studies reported higher right ear score even while attending to the left ear (Hugdahl et al., 1998; Kershner, 2016; Menashe, 2019; Morton & Kershner, 1993).

3.2.2..2. Performance difference

When compared with the control group, Most of the studies reported that LD showed poorer performance in right ear scores for attended right ear conditions (Aylward, 1984; Boliek et al., 1988; Bowen & Hynd, 1988; Hakvoort et al., 2016; Helland & Asbjørnsen, 2001; Hugdahl et al., 1998; Keefe & Swinney, 1979; Kershner & Morton, 1990; Kershner & Stringer, 1991; Martínez & Sánchez, 1999; Menashe, 2019; Morton & Kershner, 1993; Morton & Siegel, 1991; Obrzut, 1979; Obrzut et al., 1985; Obrzut et al., 1993; Yeni-Komshian et al., 1975) except four, that suggest better performance in LD group than controls (Dermody et al., 1983; Kershner, 2016; Obrzut et al., 1989; Swanson & Mullen, 1983). Only four studies had reported greater left ear scores for children in the LD group while attending only in the LE (Kershner & Stringer, 1991; Menashe, 2019; Obrzut et al., 1985; Obrzut et al., 1989) and six studies suggested good RE performance in right attended condition when compared to control group (Boliek et al., 1988; Bowen & Hynd, 1988; Hakvoort et al., 2016; Hugdahl et al., 1998; Martínez & Sánchez, 1999; Morton & Siegel, 1991).

3.3. Results of meta-analysis

3.3.1. Within group comparison

To explore the ear advantage, the pooled mean right and left ear scores of both control subjects and children with LD from various studies are compared descriptively. Traditional meta-analysis was not possible in this case because of the lack of any standard formula or software to calculate the effect size within-subject variables. However, such analysis is necessary to study the direction of ear difference in different conditions across the normal and disordered groups. Table 2, shows the pooled mean values and confidence intervals (CI) for controls and individuals with LD in free recall, attended right and attended left ears separately for CVs and digits.

Table 2. Mean values and CIs for Controls and LD.

Condition	Group	Stimulus	Right scores			Left Scores
			Pooled Mean	CI+	CI−	Pooled Mean	CI+	CI−
Free recall	Control	CV	31.34	33.54	29.15	22.38	24.31	20.45
		Digit	19.74	20.45	19.02	17.59	18.37	16.80
	LD	CV	26.22	28.61	23.84	21.45	23.47	19.43
		Digit	19.21	20.55	17.87	18.06	19.32	16.80
Directed right	Control	CV	30.60	34.57	26.63	–	–	–
		Digit	53.33	58.73	47.93	–	–	–
	LD	CV	28.75	32.59	24.90	–	–	–
		Digit	54.84	61.32	48.35	–	–	–
Directed left	Control	CV	–	–	–	22.33	25.31	19.35
		Digit	–	–	–	39.56	47.37	31.74
	LD	CV	–	–	–	25.51	28.40	22.62
		Digit	–	–	–	43	52.01	33.98

CI – confidence interval.

The pooled mean ear difference (R-L) was +8.96 in controls indicating a higher right ear score for CVs. Thus, the REA was consistent across the study as indicated by lower CI values. Similar findings were seen in children with learning disabilities too. However, the ear difference was smaller when compared to controls. Further, such large ear differences were not seen for digits in both groups (CG = 2.15 and LD = 1.15). This may be due to saturation scores that are commonly seen for dichotic digit tests.

A similar comparison in the directed attention condition shows, significantly greater RE scores in the control group for both CVs (13.26) and digits (25.96). However, such a difference was seen only for digits in LD (17.21) and not for CVs (0.03).

3.3.2. Performance difference

Meta-analysis was done with 14 studies that have been included in the current review (Boliek et al., 1988; Bowen & Hynd, 1988; Gupta, 2015; Helland et al., 2008; Hugdahl, 2003; Hugdahl et al., 1998; Hynd et al., 1979; Kershner & Micallef, 1992; Koomar & Cermak, 1981; Moncrieff & Black, 2008; Obrzut et al., 1981; Obrzut et al., 1985; Obrzut et al., 1989). The results showed an overall effect size of Z = 2.78 which was statistically significant (p = 0.005). The weighted mean value is 2.46 with a CI of 0.73, 4.19. The summary measure obtained shows controls has a larger ear difference than the LD group. The I2 value calculated is 100%, reporting highly heterogeneous data (Figure 2).

Figure 2. Forest plot indicating the performance difference between controls and LDs in dichotic CV test.

Similarly, a meta-analysis for the dichotic digit's test was done with nine studies (Aylward, 1984; Beaumont et al., 1981; Hakvoort et al., 2016; Keefe & Swinney, 1979; Kershner et al., 1984; Koomar & Cermak, 1981; McKeever & VanDeventer, 1975; Moncrieff & Black, 2008; Pinheiro et al., 2010) with available data reported a statistically insignificant summary measure with effect size, Z = 0.08 (p = 0.94) suggesting no difference between the groups. The obtained pooled mean value is −0.08 with –2.16, 1.99 CI. The I2 value is 99%, which suggests heterogeneous data. The funnel plots obtained, suggested no significant publication bias for all the conditions (Figure 3).

Figure 3. Forest plot indicating the performance difference between controls and LDs in dichotic digit test.

3.3.3. Right ear performance

In this, the mean right ear scores across the studies were compared between the two groups (LD and controls). This meta-analysis included 14 studies with dichotic CV and 9 studies with dichotic digits (same as in the above section). The results showed an effect size of Z = 2.93 which was statistically significant (p = 0.003) and suggested larger right ear scores for CVs in the control group. The pooled mean value is 3.04 for the CI 1.01, and 5.08, and the data was found to be highly heterogeneous with an I2 value of 89%. A similar comparison using dichotic digits also showed identical results, i.e., larger RE performance in the control group. The effect size was Z = 1.91 with p = 0.06 and I2 = 99%. The pooled mean value is 3.60 and CI −0.09, 7.00 (Figures 4 and 5).

Figure 4. Forest plot indicating the right ear performance difference between controls and LDs in dichotic CV test.

Figure 5. Forest plot indicating the right ear performance difference between controls and LDs in dichotic digit test.

3.3.4. Left ear performance

The mean LE scores across studies were compared between the LD and control groups. The results for CVs obtained an overall effect size of Z = 1.59 which is statistically insignificant (p = 0.11) and an I2 value of 45% indicating moderate heterogeneity in the data. The pooled mean was 0.63 for the CI −0.15, 1.41. Similarly, for digits the effect size was Z = 1.83, insignificant (p = 0.07), I2 = 99% and pooled mean value of 3.53 for CI −0.26, 7.33 was obtained. Hence no significant group difference was derived for CVs and digits (Figures 6 and 7).

Figure 6. Forest plot indicating the left ear performance difference between controls and LDs in dichotic CV test.

Figure 7. Forest plot indicating the left ear performance difference between controls and LDs in dichotic digit test.

3.3.5. Right directed attention

The mean scores for the attended right ear were compared between the two groups (Boliek et al., 1988; Bowen & Hynd, 1988; Dermody et al., 1983; Helland & Asbjørnsen, 2001; Hugdahl et al., 1998; Kershner, 2016; Kershner & Micallef, 1992; Morton & Kershner, 1991; Obrzut et al., 1985; Obrzut et al., 1989) and seven studies with dichotic digits (Hakvoort et al., 2016; Kershner & Micallef, 1991; Kershner & Stringer, 1991; Moncrieff & Black, 2008; L. Morton & Kershner, 1991, 1993; Morton & Siegel, 1991) were included in this meta-analysis. The results for CVs showed a pooled mean value of 2.81 for the CI 0.63, 4.98, and the overall effect size, Z = 2.53 (p = 0.01) suggesting a significant group difference in the RE performance for the attended RE condition. I2 value was 92% suggesting heterogeneous data. Similar findings were seen concerning dichotic digits too. Z = 2.88 (p = 0.004) and pooled mean value 2.92 for the CI 0.93, 4.91 was obtained. An I2 value of 62% suggests less heterogeneous data compared to other results (Figures 8 and 9).

Figure 8. Forest plot indicating the right directed attention performance difference between controls and LDs in dichotic CV test.

Figure 9. Forest plot indicating the right directed attention performance difference between controls and LDs in dichotic Digit test.

3.3.6. Left directed attention

The mean attended left ear scores of 17 studies were compared between LD and control groups. There were 11 studies with dichotic CVs included in this meta-analysis (Boliek et al., 1988; Bowen & Hynd, 1988; Dermody et al., 1983; Helland & Asbjørnsen, 2001; Hugdahl et al., 1998; Kershner, 2016; Kershner & Micallef, 1992; Morton & Kershner, 1991; Obrzut et al., 1985; Obrzut et al., 1989; Smith & Griffiths, 1987). The overall effect size, Z = 0.14 statistically insignificant (p = 0.89), pooled mean value 0.06 for CI −0.81, 0.93 was resulted. The data is heterogeneous as the I2 value is 81%. Similarly, six studies were included with dichotic digits (Hakvoort et al., 2016; Kershner & Micallef, 1991; Kershner & Morton, 1990; Kershner & Stringer, 1991; Morton & Kershner, 1993; Morton & Siegel, 1991) and resulted in a pooled mean of 1.61 for CI −0.71, 3.93 with statistically insignificant effect size, Z = 1.36 (p = 0.17). The measure of heterogeneity, I2 (51%) suggests comparatively lesser heterogeneous data. Hence no group difference was reported in both the groups for CVs and digits (Figures 10 and 11).

Figure 10. Forest plot indicating the left-directed attention performance difference between controls and LDs in dichotic CV test.

Figure 11. Forest plot indicating the left-directed attention performance difference between controls and LDs in dichotic digit test.

4. Discussion

Dichotic stimulation typically results in asymmetrical scores in the right and the left ear, reflecting the contralateral hemisphere's involvement in processing the speech stimulus. Dichotic tests are the commonly used test either in isolation or as part of a test battery like SCAN in the assessment of individuals with APD (Auditory Processing Disorder) and LD. The current review and meta-analysis are mainly focused on determining the dichotic test results in children with LD including ear asymmetry and, in turn, cortical asymmetry. Meta-analysis was done independently on two common variants of the dichotic test (digits and CVs) and two different paradigms (free recall and directed attention). The study results are discussed in the following sections.

4.1. Ear asymmetry in free recall

The first and foremost question is whether individuals with LD lack ear/hemispheric asymmetry. The overall results suggest that the LD has an ear asymmetry, but the ear difference magnitude is lesser than the control. These findings suggest that the LD has poor or reduced but is not abnormal lateralization. The forest plot (Figure 2) also suggests many studies supporting this fact. The performance difference between the ears in children with LD is so low that 4/13 studies reported no observable difference. (Boliek et al., 1988; Hugdahl et al., 1998; Kershner & Micallef, 1992; Obrzut et al., 1989).

The pooled mean difference between the ears using digits was far too small for both controls and children with LD (Table 2). Even though the pooled mean difference scores resulted in statistical significance, 4/9 studies have concluded that there was no difference in performance between LDs and controls concerning ear asymmetry (Hakvoort et al., 2016; Keefe & Swinney, 1979; McKeever & VanDeventer, 1975; Pinheiro et al., 2010). Hence, this minimal difference may be statically significant in the current review but may not be clinically significant (Figure 3). This could be due to the ceiling effect that could be commonly seen in dichotic digits and it could be reduced by increasing the digit pair length (Koomar & Cermak, 1981). Partially this lack of difference could be also explained by the wider age range of participants and heterogeneity in the dyslexics themselves (Keefe & Swinney, 1979; McKeever & VanDeventer, 1975; Pinheiro et al., 2010).

4.2. Performance asymmetry in free recall

The above section explores the different scores obtained between the ears (Right-Left) in LD and controls. However, it could be possible that the individuals with LD underperform overall in both ears yet preserve the ear difference. Hence, it is interesting to compare the absolute mean scores of right ears between these groups for each ear independently. The overall results of the current review suggest that there is a significant difference in the right ear scores between the control and the LDs (Table 2), forest plot (Figure 4). However, on closer observation, we see that most studies in the forest plot have their confidence interval crossing the line of no effect and the mean difference observed in the individual studies was not greater than 3 scores. Except for three recent studies which report very poor right ear scores in LD compared to controls. Hence overall, it can be considered that there may not be any clinically significant difference between the absolute scores of right ears. Interestingly, two studies also observed an inverse result, where LDs performed better (Boliek et al., 1988; Dermody et al., 1983) said that this better performance by the right ear in individuals with LD is due to their inability of allocating processing resources to a difficult listening task when compared to their controls. A comparison of absolute left ear scores (Figure 6) also showed no significant difference between the groups, though the variability and trend were like that of the right ear. Studies using dichotic digits also showed no significant absolute scores difference between LD and control groups (Figures 5 and 7).

4.3. Performance differences in directed attention

The overall results suggest a statistically significant difference between children with LD and controls in right attended condition but not in the left-attended condition. Table 2 and Figure 8 show that most studies agree with this finding. Adding a constrain of attention to the dichotic listening makes the task cognitively demanding since it recruits higher cognitive resources to perform this task. However, attending left and right refers to two different cognitive tasks as such. The right ear-focused attention condition is considered a selective attention task since the right is considered the naturally dominant for language processing and the individual have to simply ignore the left. However, attending to the left is more of an executive function task as they have to suppress the dominant right ear information to be perceived. In the current review, there is any evidence that the LD suffers selective attention deficit and not executive attention. However, Figure 8, and Table 2 show two grossly deviant results in the literature. Overall results suggest that controls score better in right attended condition than the left, but the children with dyslexia scored equally in both these conditions (Helland, 2001). This lack of right attended advantage is absent in individuals with a learning disability. However, they did not discuss these findings in light of the cognitive processes that are involved in these two conditions but rather accounted for the language deficits associated with LD. The opposite findings (Obrzut et al., 1989) suggest that such asymmetry in the attended condition is seen only in individuals with LD and not in controls. They argue this ability to modulate attention is inhibited in typically developing children due to hard-wired language preference, while LDs lack this. However, their findings by them have not been replicated since 1989. Similar results were obtained for the dichotic digit test for directed right (Figure 9) and left (Figure 11) conditions, which reiterates the findings from dichotic CVs.

Overall, one can see poor performance in both free recall and directed attention conditions. Free recall using CVs and attended right ear scores were sensitive to produce significant differences between LD and control children. In most instances where there was a statistically significant difference, the magnitude was relatively less (eg: Figure 8). It is important to infer the variations in test material and scoring patterns before inferring this difference. Dichotic tests are classical CAPD tests that can be developed in-house and hence, in most of these studies, the authors have developed an indigenous test and scoring pattern and administered them to the population. Hence, though all the articles included in the review and the test scores varied across studies stating from 12 to 1000+ which could influence the clinically significant difference. Further, no dichotic tests have provided their psychometric functions or sensitivity/specificity for detecting individuals with LD or any other clinical population. or any other clinical populations. Also, the current study included only articles published till 2020 and thus missing any novel studies published after 2020. Further subgroup analysis of pooled data based on important outcomes like handedness, age and hearing status could not be done which could be considered for future research.

5. Conclusion

The REA is preserved in children with LD for the dichotic free recall test. The right ear scores were significantly poor compared to controls. There was no evidence of LEA in children with LD. Dichotic CVs represented these differences better than digits. These results suggest, poor lateralization for language. In the directed attention condition, the children with LD showed difficulty in right attended conditions whereas no deficit is observed in left attended conditions. These results suggest a deficit in selective attention than executive attention. No difference was seen between dichotic CVs and digits in directed attention conditions.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data is available with the corresponding author and will be shared on request.