1）Keisuke Miyazak , 2） Yosuke Hashimoto,3） Hitoshi Uchiyama, 4）Miwa Sakai
Abstract
We conducted a learning program using haptic reading plates with raised characters of specific heights, widths, and shapes for eight children with dyslexia (all 8 had difficulties with RAN, all 8 had reading difficulties, and 4 out of 8 had writing difficulties). The training involved “reading aloud while touching and looking at” characters, focusing on decoding practice with kana (hiragana and katakana), special syllables, words, and short sentences.
After the training, 7 out of 8 children showed reduced time on the RAN tasks in the STRAW-R, and 7 out of 8 also demonstrated shorter reading times on rapid reading tasks. Additionally, all 4 children with difficulties in writing kana improved their accuracy in dictation tasks. Furthermore, children who showed improvements in the STRAW-R results reported a subjective reduction in the burden of reading and writing in a questionnaire.
These results suggest that haptic learning facilitated by active touch promoted the formation of character shape memory and the associative memory between characters and sounds. However, the learning tasks in this study did not include numbers or pictures, which are used in RAN tasks, making it difficult to directly attribute the improvements in naming speed to the learning program itself. Nevertheless, it is hypothesized that the enhanced recall of character shape memory developed through haptic reading served as a cognitive cue in working memory, thereby facilitating the efficient retrieval of phonological information.
Key word: Haptic Learning, Developmental Dyslexia, STRAW-R，RAN
1) Speech-Language-Hearing Therapist, Miyazaki Speech Language Therapy Co., Ltd.
TEL : 050-3556-3047 e-mail : sawaru126＠gmail.com
Address: 2F Mitsubayashi Building, 520 Higashi Shiokoji-cho, Shichijo Sagaru, Higashinotoindori, Shimogyo-ku, Kyoto600-8174, Japan
2) Assistant Professor, Japanese Language Programme, Faculty of International Liberal Arts, Akita International University
TEL : 018-886-5929 e-mail : yosuke@aiu.ac.jp
Address: Okutsubakidai-193-2 Yuwatsubakigawa, Akita, 010-1211, Japan
3) Associate Professor, The University of Shimane
TEL：0855-24-2200 e-mail : h-uchiyama@u-shimane.ac.jp
Address: 7 Chome-24-2 Hamanogi, Matsue, Shimane 690-0044, Japan
4) Speech Pathologist, Tandem Allied Health
TEL: 0402-601-537 e-mail : miwa.sakai.1995@gmail.com
Address: 1 Waddell St, Bacchus Marsh 3340 Victoria, Australia
Introduction
Children with no delays in general intellectual development and no sensory impairments such as in vision or hearing, yet who show significant delays in reading and writing abilities, are recognized in the fields of medicine and education as having “Developmental Dyslexia”1).
These children experience difficulties in fluent reading and in acquiring and expressing written forms of characters. Regarding “reading,” one primary cause of difficulty is weakness in phonological processing, which results in poor decoding of the correspondence between letters and sounds. Additionally, a weakness in visually recognizing strings of letters as unified word forms also hinders fluent reading.Research on the neurological basis of this condition has shown that decoding letters and sounds primarily involves the left temporoparietal region, with the left inferior frontal gyrus playing a supportive role. Furthermore, the visual recognition of word forms involves the left fusiform gyrus2). These reading difficulties often cause mental fatigue, potentially lowering academic motivation3).Regarding “writing,” weaknesses in visual recognition and visuospatial memory affect the learning and expression of complex character forms like kanji4).
Background Research and Instructional Methods
Research by Uno et al. has revealed that developmental dyslexia involves multiple cognitive factors, including not only phonological processing but also visual recognition and naming speed5). Notably, slower naming speed before school entry has been identified as a predictor of later reading difficulties6).
In Japan, instructional methods to improve reading include decoding instruction centered on oral reading. This approach has been shown to enhance reading fluency by teaching the correspondence between letters and sounds, and gradually improving word form recognition 7). For kanji learning, the “bypass method” (auditory-based learning), which emphasizes verbalizing the structure of kanji to aid memory, has been found effective9).
Internationally, multisensory learning incorporating haptic exploration has been applied to children with reading and writing difficulties. For example, studies have demonstrated the effectiveness of using clay to form alphabet letters10). haptic -based learning has also been applied to preschool children, where it was found to improve their understanding of letter-sound correspondence compared to visual exploration alone 11).
Objectives of This Study
Based on the benefits of haptic feedback and dual coding, we developed a multisensory learning method utilizing 3D haptic-readable character plates. This method involves touching characters while reading aloud, using haptic stimuli as cognitive cues to enhance memory of character forms, association between letters and sounds, and word form recognition. In this study, we conducted a pilot experiment with eight children with reading and writing difficulties to verify the feasibility of this approach before applying it in a controlled group comparison.
Ⅰ. Methods
1. Participants
This study involved eight upper-grade elementary school children (mean age: 10.4 years, range: 9.3–12.4 years) who had no visual or auditory impairments but exhibited difficulties in reading and writing kana. Among them, five children had been medically diagnosed with learning disabilities (LD), and two of these also had comorbid developmental disorders such as ADHD or ASD. The remaining three children, though undiagnosed, demonstrated significant difficulties in reading and writing as reported by their guardians and based on the discrepancy between their intelligence and literacy test results, fulfilling the criteria for developmental dyslexia.
2. Evaluation of Reading and Writing Abilities
The participants’ abilities in reading and writing kana were assessed using the Revised Standardized Reading and Writing Screening Test (STRAW-R). The test included the following tasks:
• Rapid Reading Task: Measured the time required to read five types of items (hiragana words, non-words, katakana words, non-words, and sentences) to evaluate fluency.
• Dictation Task: Assessed the accuracy of writing 20 kana characters (hiragana and katakana) dictated aloud.
Based on the STRAW-R manual, a z-score of 1.5 or above in reading time was considered indicative of “reading difficulty,” while a z-score of 1.5 or below in writing accuracy indicated “writing difficulty.” All eight participants were classified as having “reading difficulty,” and four (Cases 2, 3, 6, and 7) also exhibited “writing difficulty” (Table 1).
3. Evaluation of Intellectual Levels and Cognitive Functions
To assess overall intellectual levels, the Raven’s Colored Progressive Matrices (RCPM) test was administered. The participants’ mean total score was 34 points (mean z-score: 0.5), confirming that all participants were within the average range of intellectual ability.
Cognitive functions were evaluated based on three components:
1. Rapid Automatized Naming (RAN): Measured using STRAW-R tasks requiring alternating naming of pictures and numbers.
2. Visual Cognitive Function (V): Assessed through the Rey-Osterrieth Complex Figure Test (ROCFT), including copying and delayed recall tasks.
3. Phonological Awareness (P): Evaluated using tasks involving repetition and backward repetition of non-words.
Participants with z-scores of 1.5 or higher in these tests were classified as having difficulties. All eight participants exhibited issues with RAN (R), while visual cognition (V) problems were noted in Cases 5 and 8, and phonological awareness (P) problems in Cases 3, 4, and 8 (Table 2).
4. Psychological Burden and Academic Performance
Psychological burden related to “reading” and “writing” was assessed through surveys given to both the children and their guardians. All eight participants reported a burden in these areas. Academic performance in Japanese language was categorized into four levels: severe delay of more than two grades (1 child), delay of one to two grades (4 children), below-average performance (2 children), and no delay (1 child).
Ethical Considerations
This study was conducted following approval by the ethics committee of Akita International University (Approval No. 0217).
Table 1: Profiles of Eight Children’s Symptoms and Results of RAN and STRAW-R Tasks Before Haptic Reading Training
Based on a cutoff z-score of 1.5, a “*” was assigned to indicate areas of difficulty. For the RAN and rapid reading tasks, difficulty was determined when the time required exceeded a z-score of 1.5. For the dictation task, difficulty was identified when the number of correct responses fell below a z-score of -1.5. “Hira” refers to hiragana, and “Kata” refers to katakana.
Table 2. Evaluation of Cognitive Functions and Type Classification of Eight Children
Based on a cutoff z-score of 1.5, a “*” was assigned to indicate areas of difficulty. Children with a z-score of 1.5 or higher in RAN (Rapid automatized naming) task time were classified as “R.” Children with a z-score of 1.5 or below in either the 3-minute or 30-minute delayed recall drawing task of the ROCFT (Rey-Osterrieth Complex Figure Test) were classified as “V” (Visual cognitive function). Children with a z-score of 1.5 or below in the reverse repetition task of 4-mora words or non-words were classified as “P” (Phonological function).
2. Haptic Reading Tasks Using 3D letter plates
Figure 1. Method for Creating haptic Reading Characters
The 3D data of the hiragana word “みかん” (bottom) and the haptic character (top) produced using a 3D printer. The haptic character has a height of 0.5 mm, a width of 3.0 mm, and a rounded convex shape, with the raised parts painted black.
The haptic reading plates consisted of 3D-raised kana characters, words, and short sentences. These haptic characters were created using 3D data modeled with Autodesk Fusion360 and fabricated with a 3D printer that embossed the character outlines (Figure 1). The haptic reading plates were formatted into A4-sized sheets as follows:
1. Kana Characters (Unvoiced, Voiced, Semi-voiced)
o Four plates containing 71 hiragana and 71 katakana characters, each embossed at 62 pixels per character.
2. Special Syllables (Palatalized and Nasal Sounds)
o Two plates of 27 hiragana morae and 27 katakana morae, with characters embossed at 62 and 32 pixels.
3. Kana Words
o Six plates of 68 hiragana words and 68 katakana words, each word consisting of 2 to 7 morae, with characters embossed at 45 pixels.
4. Kana Sentences
o Two plates containing 120-character sentences embossed at 32 pixels per character.
A total of 14 plates were created, with character embossing specifications based on prior memory experiments using ROCFT (Rey-Osterrieth Complex Figure Test). The embossed areas were 0.5 mm in height, 0.3 mm in width, and rounded at the edges, with black coloring applied to the raised portions for visual and haptic recognition.
3. Evaluation and Learning Methods
1. Evaluation
Before starting the learning tasks, an initial evaluation (first session) was conducted using STRAW-R tasks, including RAN, rapid reading, and single kana dictation (writing from dictation) to assess the children’s reading and writing abilities. After completing the haptic learning tasks over 40 days (14 sessions), a second evaluation was conducted 40 days after the initial assessment. The same STRAW-R tasks were repeated, and changes between the first (pre-learning) and second (post-learning) results were compared to examine the learning effect.
Additionally, on the day of the second evaluation, children completed a custom questionnaire to assess changes in psychological burdens related to reading and writing. The questionnaire included items on the “burden of reading text” and “burden of writing text,” with three options: “burden decreased,” “burden increased,” or “no change,” along with reasons for their answers.
2. Learning Tasks and Haptic Reading Methods
The learning approach combined face-to-face instruction based on decoding guidance with monitoring support. At the first session, both children and their parents received instruction on how to proceed with haptic reading (e.g., methods of touching, number of sheets per session, and suggested time per sheet). Subsequent sessions alternated between home practice and monitored feedback.
The learning tasks involved completing all 14 haptic reading plates using the method of “reading aloud while touching and observing.” The plates were divided into the following four sections, completed one section per day:
• ① Kana Characters (Hiragana and Katakana): 4 plates of unvoiced, voiced, and semi-voiced kana characters.
• ② Special Syllables (Palatalized and Nasal Sounds): 2 plates of hiragana and katakana.
• ③ Kana Words: 6 plates of hiragana and katakana words.
• ④ Kana Sentences: 2 plates of short sentences.
All 14 plates constituted one session, with a total of 14 sessions completed over 40 days. Individual differences in learning pace were accommodated; children who could not complete all 14 plates in one day spread their sessions over multiple days.
During the tasks, specific haptic reading techniques were instructed:
• For single kana characters, children traced each character with their fingertip for at least two seconds while reading aloud.
• For special syllables (e.g., “nya,” “ja,” “pya”), children traced the sequence of characters as a unit while reading aloud.
• For kana words (2–7 morae) and sentences, children traced the entire word or phrase while observing and reading aloud.
Midway through the 40-day period, face-to-face monitoring sessions were conducted to maintain motivation, review proper techniques, and provide feedback. After completing the learning tasks, the second evaluation was conducted. Parents were also provided with a log sheet to record the learning sessions, including dates, duration, and the number of plates completed.
Ⅱ. Results
STRAW-R results of 8 children
Table 3 presents the z-score differences for writing tasks, speed-reading tasks, and RAN tasks in the first session (pre-training) and the second session (post-training), as well as the results of subjective evaluations.
Table 3. Changes in STRAW-R Performance (z-score differences) and Subjective Evaluations After Haptic Reading Learning
Scores with a z-score improvement of 1.0 or higher are marked with an asterisk (*). For RAN and speed-reading tasks, improvements are indicated by reduced time (−), while for dictation tasks, improvements are reflected in the number of correct responses (+).
1. Writing Tasks in Hiragana and Katakana (STRAW-R)
All four children (Cases 2, 3, 6, 7) who struggled with writing single kana characters improved in at least one of the two items, with a z-score difference of 1.5 or more. None of the children showed a decline exceeding a z-score difference of 1.0 across all items. Notably, Case 7 demonstrated significant improvement in both items, with a z-score increase of 1.5 or more. Across all four children, the number of characters with delayed responses exceeding 2 seconds from hearing to writing decreased from 8 to 4. In the subjective evaluation, three children reported “writing has become easier,” while Case 3 reported “no change.”
2. Speed-Reading Tasks (STRAW-R)
Seven out of the eight children (Cases 1, 2, 3, 4, 6, 7, 8) who had difficulty reading demonstrated a z-score difference of 1.5 or more in reading time reduction in at least one of the five items. With a threshold of z-score difference 1.0, all seven showed time reductions in two or more items. There was no significant difference between the changes in scores for word and non-word items. None of the children exhibited delays exceeding a z-score difference of 1.0 across all items. Case 7 showed the most improvement, with four items achieving a z-score difference of 2.0 or more. Case 5, with the least improvement, demonstrated a z-score difference of 1.0 or more in only one item (sentence reading). Subjective evaluations revealed that the seven children with significant time reductions reported “reduced reading burden,” while Case 5 reported “no change.”
3. RAN Tasks (STRAW-R)
Seven out of eight children (Cases 1, 2, 3, 4, 6, 7, 8) who struggled with RAN tasks showed reduced average time across the three RAN tasks. Cases 1, 2, 3, 4, 6, and 8 achieved time reductions with a z-score difference of 1.5 or more, while Case 7 had a z-score difference of 1.0 to 1.5. Case 5 did not show significant time reductions.
4. Analysis Based on Cognitive Background Functions
We analyzed the results considering the cognitive processes involved in multisensory learning. The reminiscence effect, known to enhance learning outcomes through the integration of visual, motor, and haptic information over time, was considered relevant to this study. In ROCFT performance changes, children with a z-score difference of -1.0 or lower in delayed recall performance were identified as having difficulty retaining multisensory memory (Cases 5 and 8). On the other hand, Cases 1, 2, and 7 showed characteristics of easy retention of multisensory memory. Notably, Case 7 demonstrated the highest number of improved items across both speed-reading and writing tasks.
5. Observed Changes During Task Implementation
Through face-to-face instruction and monitoring over 20 days, some children (Cases 1, 2, 6) became capable of haptic reading with their eyes closed, limited to single kana characters. While the task required visual and haptic integration, these changes were observed incidentally when the children attempted to identify kana characters by touch alone. Sequential reading evolved into fluent reading for words in Cases 1, 3, 4, 6, and 7. Additionally, exploratory touch with the finger pads replaced tracing the kana lines with fingertips in Cases 2 and 8. No such changes in touch methods were observed in Case 5.
III. Discussion
The observed improvements in the accuracy of writing tasks and reduced reading time in speed-reading tasks were analyzed from the perspective of “active touch.” Previous memory experiments utilizing 3D representations of ROCFT diagrams, similar in shape and height to the haptic reading board used in this study, have shown that combining visual observation with haptic exploration significantly enhances recall performance. This improvement is attributed to active touch directing attention toward the morphological features of complex shapes, which facilitates memory formation and recall. Improvements in Writing Performance In this study, all four children (Cases 2, 3, 6, and 7) who demonstrated improved accuracy in kana writing tasks likely benefited from touching enlarged, 3D kana characters (65 pixels per character). This haptic interaction may have helped form more concrete and detailed memory representations of character shapes.
Furthermore, the number of delayed responses exceeding two seconds before writing decreased from 8 characters before the intervention to 4 afterward. This suggests an improvement in the ability to recall character shapes, which likely contributed to subjective evaluations where three out of the four children reported that “writing has become easier.” Additionally, the haptic memory representations formed through this process may have enabled the children to recognize single kana characters by touch alone, even with their eyes closed. Promotion of Reading Fluency In the speed-reading tasks, seven of the eight children (Cases 1, 2, 3, 4, 6, 7, and 8) showed reduced reading times. It is hypothesized that the character shape memories formed through haptic learning facilitated top-down visual cognitive processing, thereby promoting more efficient character recognition. Previous studies have reported similar findings, such as improved performance in decoding tasks through haptic exploration of 3D alphabetic characters. These studies suggest that haptic perception integrates simultaneous visual and sequential phonological information, which strengthens the connection between spelling and pronunciation, thereby enhancing memory traces. In this study, reading aloud while receiving haptic feedback from 3D kana characters may have efficiently facilitated the formation of associative memories between characters and sounds. The learning tasks used (haptic reading board) excluded kana words overlapping with the speed-reading tasks in STRAW-R but still demonstrated reading time reductions for both meaningful words and non-words. This finding suggests that the improvements were not merely due to changes in word-form recognition but rather resulted from enhanced recognition of individual kana shapes and the formation of character-sound associative memories. These associative memory formations likely improved decoding efficiency in the speed-reading tasks and enhanced the encoding process of recalling characters from phonological information in the writing tasks. Monitoring during the intervention revealed a shift from “sequential reading,” where children touched and read each kana character individually, to “chunk reading,” where they read entire words as cohesive units. This change may have been facilitated by increased alertness through haptic perception, which promoted memory chunking—the process of integrating individual character information into word-level representations.
Haptic Learning and Cognitive Characteristics Further analysis focused on the retention characteristics of the children, using performance changes in the ROCFT delayed recall task (3 minutes vs. 30 minutes). Among the children with positive performance changes, three included the most improved case (Case 7), while two of the children with negative changes included the least improved case (Case 5). In Case 5, weaknesses in central integration functions, necessary for constructing whole images from partial information, were evident. This may have hindered the effectiveness of haptic learning for forming and recalling character shape memories. On the other hand, Case 8, despite showing weaker memory retention characteristics, demonstrated improvements in two speed-reading items and the RAN task. This indicates that memory retention characteristics alone cannot fully account for the effects of haptic learning. Further studies using haptic memory tasks more closely aligned with haptic reading activities are needed to verify these findings.
Promotive Changes in RAN In the RAN tasks, seven of the eight children exhibited reduced naming times. RAN tasks require the ability to efficiently recall phonological information while visually recognizing targets and shifting attention, which are complex cognitive functions. The haptic reading tasks in this study did not directly include stimuli used in the RAN tasks, such as numbers and pictures, but the improvements in RAN may be attributed to the reinforcement of character shape memories through haptic learning. The recall of language-related information involves two routes: the “naming route,” where words are recalled from phonological dictionaries via semantic memory, and the “writing route,” where words are recalled from visual dictionaries. In cases of naming difficulty, providing visual cues of the initial letter has been reported to improve naming performance by facilitating the interaction between visual and phonological dictionaries. The mechanisms underlying the improvements in RAN performance may include enhanced recall of character shape memories facilitated by haptic learning, which promoted the retrieval of phonological information from visual dictionaries. The process of “seeing while touching and reading aloud” required during the haptic learning tasks likely enhanced multisensory integration, linking visual, haptic, and auditory information. This multisensory approach may have strengthened the connection between semantic memory, visual dictionaries, and phonological dictionaries.
Considerations for Brain Function Studies Looking forward, brain function studies can provide further insights into these mechanisms. Previous research on the developmental trajectory of naming abilities has shown that improvements in naming are accompanied by increased connectivity between the left fusiform gyrus and other brain regions involved in memory encoding and phonological processing, such as the left parahippocampal gyrus and left inferior frontal gyrus. The left fusiform gyrus plays a critical role in integrating visual and haptic information and serves as a visual dictionary area for recalling character shapes and recognizing word forms during reading and writing. The multisensory reading activities facilitated by haptic learning may influence changes in connectivity within this network, centered around the left fusiform gyrus. However, the precise mechanisms promoting faster naming and improved reading fluency remain unclear and warrant further investigation using functional MRI or other neuroimaging techniques. While this study is based on a small sample of eight cases, future research should include larger comparative studies between intervention and non-intervention groups, along with measurements of brain function changes following haptic learning, to provide a more detailed understanding of its effects and mechanisms.
References
1. Akira Uno. (2016). Developmental Dyslexia. Higher Brain Function Research, 36(2), 170-176.
2. Ayumi Seki. (2009). Functional Imaging of Children with Developmental Dyslexia. Cognitive Neuroscience, 11(1), 54-58.
3. Tatsuya Koeda. (2020). Diagnosis, Treatment, and Early Support for Children with Dyslexia—Theory and Practice of T-Style Hiragana Oral Reading Support. Bulletin of Meisei University Developmental Support Research Center, March Issue (5), 56-58.
4. Takashi Goto, Akira Uno, Noriko Haruhara, Masato Kaneko, Tokuko Awaya, Junko Kozuka, Akiko Katano. (2010). Visual Function, Visual Perception, and Visual Cognitive Function in Children with Developmental Dyslexia. Phonetics and Speech Language Medicine, 51(1), 38-53.
5. Akira Uno, Noriko Haruhara, Masato Kaneko, Tokuko Awaya, Junko Kozuka, Takashi Goto. (2018). Cognitive Deficits Underlying Developmental Dyslexia: A Comparison with Age-Matched Control Groups. Higher Brain Function Research, 38(3), 267-271.
6. Masato Kaneko, Akira Uno, Noriko Haruhara, Tokuko Awaya. (2012). Predicting Reading Difficulties After School Entry in Preschool Children: The Utility of Rapid Automatized Naming as a Screening Test. Brain and Development, 44(1), 29-34.
7. Hitoshi Uchiyama, Daisuke Tanaka, Ayumi Seki, Eiji Wakamiya, Noriko Hirasawa, Naotaka Ikeya, Ken Kato, Tatsuya Koeda. (2013). Effects of Decoding Instruction for Children with Difficulties in Hiragana Oral Reading. Brain and Development, 45(3), 239-242.
8. Panagiotis G. Simos, Jack M. Fletcher, Shirin Sarkari, Rebecca L. Billingsley, Carolyn Denton, Andrew C. Papanicolaou. (2007). Altering the Brain Circuits for Reading Through Intervention: A Magnetic Source Imaging Study. Neuropsychology, 21(4), 485-496.
9. Akira Uno, Noriko Haruhara, Masato Kaneko, Takashi Goto, Tokuko Awaya, Junko Kozuka. (2015). Kana Training Using the Bypass Method for Children with Developmental Dyslexia—A Case Series Study on Methods and Effects of Training Adapted to the Structure of Dyslexia. Phonetics and Speech Language Medicine, 56(2), 171-179.
10. Philip A.P., Cheong L.S. (2011). Effects of the Clay Modeling Program on the Reading Behavior of Children with Dyslexia: A Malaysian Case Study. The Asia-Pacific Education Researcher, 20(3), 456-468.
11. Florence B., Edouard G., Pascale C.L., Sprenger C. (2004). The Visuo-Haptic and Haptic Exploration of Letters Increases Kindergarten Children’s Reading Acquisition. Cognitive Development, 19(3), 433-449.
12. Keisuke Miyazaki, Sumie Yamada, Sota Kawasaki. (2023). The Promoting Effect of Multisensory Learning Using Vision and Touch on Visual Memory in the Rey-Osterrieth Complex Figure Test. Cognitive Neuroscience, 24(3-4), 87-92.
13. Yurie Nishino, Hiroshi Ando. (2008). Brain Mechanisms of Object Recognition Based on 3D Shapes. Psychological Review, 51(2), 330-346.
14. Akira Uno, Naoko Niie, Noriko Haruhara, Masato Kaneko. (2017). Revised Standard Reading and Writing Screening Test (STRAW-R) —Evaluation of Accuracy and Fluency. Interna Publishing: Tokyo.
15. Akira Uno, Naoko Niie, Noriko Haruhara, Masato Kaneko. (2005). Raven’s Colored Progressive Matrices Test for Typically Developing Children—For Screening Learning Disabilities and Childhood Aphasia. Phonetics and Speech Language Medicine, 46(3), 185-189.
16. Yuko Ogibu, Sota Kawasaki, Tomohito Okumura, Makoto Nakanishi. (2019). Developmental Process and Scale Construction of the Rey-Osterrieth Complex Figure Test in Childhood. Journal of Biomedical Fuzzy Systems Association, 21(1), 69-77.
17. Keiko Hara. (2016). Chapter 5.5: Instruction and Support for Specific Disabilities, Focused on Learning Disabilities—Developmental Dyslexia. In: Hiroyo Ishida, Ikuyo Ishizaka (Eds.), Language Developmental Disorders for Speech-Language Pathologists, 2nd Edition. Ishiyaku Publishers: Tokyo, 172.
18. Norio Ando, Shuji Hasegawa. (2015). Reminiscence and Learning Effects: Insights from an English Learning Program for Elementary School Children. Journal of Uekusa Gakuen University, 7, 25-35.
19. Chihiro Wakamatsu, Sumio Ishiai. (2018). Mechanisms of Naming Facilitation by Initial Kana Letter Cues: A Study on Aphasic Cases with Weak Initial Phoneme Cue Effects. Neuropsychology, 34(4), 299-309.
20. Ravindra Arya, Brian Ervin, Jason Buroker, Hansel M. Greiner, Anna W. Byars, Leonid Rozhkov, Jesse Skoch, Paul S. Horn, Clayton Frink, Craig Scholle, et al. (2022). Neuronal Circuits Supporting Development of Visual Naming Revealed by Intracranial Coherence Modulations. Frontiers in Neuroscience, 16, 867021.