Neurocase The Neural Basis of Cognition ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/nncs20 Selective impairment of On-reading (Chinesestyle pronunciation) in alexia with agraphia for kanji due to subcortical hemorrhage in the left posterior middle temporal gyrus Mizuho Yoshida , Toshihiro Hayashi , Kurumi Fujii , Hiroyuki Ishiura , Shoji Tsuji & Yasuhisa Sakurai To cite this article: Mizuho Yoshida , Toshihiro Hayashi , Kurumi Fujii , Hiroyuki Ishiura , Shoji Tsuji & Yasuhisa Sakurai (2020) Selective impairment of On-reading (Chinese-style pronunciation) in alexia with agraphia for kanji due to subcortical hemorrhage in the left posterior middle temporal gyrus, Neurocase, 26:4, 220-226, DOI: 10.1080/13554794.2020.1788608 To link to this article: https://doi.org/10.1080/13554794.2020.1788608 Published online: 16 Jul 2020. Submit your article to this journal Article views: 21 View related articles View Crossmark data Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=nncs20 NEUROCASE 2020, VOL. 26, NO. 4, 220–226 https://doi.org/10.1080/13554794.2020.1788608 Selective impairment of On-reading (Chinese-style pronunciation) in alexia with agraphia for kanji due to subcortical hemorrhage in the left posterior middle temporal gyrus Mizuho Yoshidaa,b, Toshihiro Hayashi a,c , Kurumi Fujiia, Hiroyuki Ishiuraa, Shoji Tsujia,d,e and Yasuhisa Sakurai f a Department of Neurology, The University of Tokyo, Tokyo, Japan; bDepartment of Rehabilitation Medicine, Showa University, Tokyo, Japan; Department of Physiology, Teikyo University, Tokyo, Japan; dDepartment of Molecular Neurology, The University of Tokyo, Tokyo, Japan; eInstitute of Medical Genomics, International University of Health and Welfare, Narita, Japan; fDepartment of Neurology, Mitsui Memorial Hospital, Tokyo, Japan c ABSTRACT ARTICLE HISTORY We report a patient with alexia with agraphia for kanji after hemorrhage in the left posterior middle temporal gyrus. The results of single-character kanji reading and two-character on- (Chinese-style pro­ nunciation), kun- (native Japanese pronunciation), and Jukujikun (irregular kun-) reading word tests revealed that the patient could not read kanji characters with on-reading but read the characters with kun-reading. We consider that this on-reading alexia was caused by disconnection between the posterior inferior temporal cortex (orthographic lexicon) and the posterior superior temporal gyrus (pho­ nological lexicon), and preserved kun- and Jukujikun-reading was realized by bypassing the orthographyto-phonology route by the semantic route. Received 24 January 2020 Accepted 18 June 2020 Introduction Nonaphasic alexia with agraphia occurs following damage to the left angular gyrus (plus the lateral occipital gyri) and the left poster­ ior inferior temporal cortex (Brodmann area 37) (Sakurai, 2011). Because of the two distinct writing systems in the Japanese lan­ guage, kanji (Japanese morphograms of Chinese origin) and kana (Japanese phonetic syllabary writing), there may be instances where Japanese patients with alexia with agraphia show dissocia­ tive disturbances between kanji and kana. For example, kana read­ ing was prominently disturbed in an angular gyrus lesion, whereas kanji reading and writing were selectively impaired in a lesion of the posterior inferior temporal cortex (Iwata, 1986). Moreover, the combination of kana alexia and kanji agraphia caused by damage to the occipital and inferior temporal gyri was reported (Mochizuki & Ohtomo, 1988). In angular alexia with agraphia, reading is more impaired in kana than in kanji, whereas writing is more impaired in kanji than in kana (Kawamura, 1990; Miyake & Tanaka, 1992). Neuropsychologically, kana alexia encompasses a letter identi­ fication deficit and a grapheme–phoneme conversion disorder, whereas kanji agraphia corresponds to lexical agraphia (selec­ tive impairment of irregular word spelling) (Ullrich & Roeltgen, 2011) in European countries. In alexia with agraphia due to a posterior inferior temporal cortex lesion, kanji reading and writing are predominantly impaired, although kana reading and writing are more or less disturbed. We designated this symptom as orthographic alexia with agraphia in the sense that errors are derived mainly from damaged visual word forms (Sakurai, 2004). There are two ways of reading kanji: on-reading (Chinesestyle pronunciation at the time of borrowing) and kun-reading CONTACT Yasuhisa Sakurai ysakurai-tky@umin.ac.jp © 2020 Informa UK Limited, trading as Taylor & Francis Group KEYWORDS On-reading; kun-reading; onreading alexia; kanji; middle temporal gyrus (native Japanese pronunciation). A kanji character usually has multiple on- and kun-readings. Both on-reading and kunreading convey the phonetic value and meaning. However, onreading has many homophones. For example, the on-reading /shin/ is shared with 29 out of 2136 daily-use kanji characters (https://www.bunka.go.jp/kokugo_nihongo/sisaku/joho/joho/ kijun/naikaku/kanji/joyokanjisakuin/). Therefore, it is difficult to identify a kanji only using a single on-reading; a kanji can be identified using kun-reading or the context. In this sense, onreading, which originally means “pronunciation of a word” in Japanese, primarily conveys the phonetic value according to the Chinese pronunciation, and thus should be called phonetic reading. In contrast, kun-reading has no or at most two homo­ phones with different accents. Thus, it is easy to identify a kanji using a single kun-reading. For this reason, kun-reading, which originally means “interpretation or meaning of a word” in Japanese, has direct access to the meaning of kanji, and thus should be called semantic reading. This dual reading system can cause on–kun distinction in some brain-damaged patients. We found that kun-reading was predominantly impaired in patients with semantic dementia. When reading a kanji character, the patients read it first with on-reading irrespective of the context and did not recall kunreading (Y. Sakurai et al., 2006a). This on-preceding and kundeletion phenomenon is characteristic of Japanese semantic dementia. Furthermore, patients have difficulty reading speci­ fic-reading words called Jukujikun. Jukujikun is the semantic assignment of Chinese characters to native Japanese words (Kindaichi, 1988): e.g., 眼鏡 ([megane], eye glasses) consists of two kanji characters 眼 ([gaN] (on-reading), [me] (kun-reading), eye) and 鏡 ([kyou] (on-reading), [kagami] (kun-reading)], mir­ ror). The pronunciation [megane] is unpredictable from those NEUROCASE of the two constituent characters. Thus, Jukujikun is regarded as irregular kun-reading (Kess & Miyamoto, 1999). Although kun- and Jukujikun-reading are predominantly impaired in patients with semantic dementia, it is not known whether there is a clinical state wherein on-reading is predo­ minantly impaired. We report a patient who showed pro­ nounced on-reading impairment with relative preservation of kun- and Jukujikun-reading following hemorrhage in the left posterior middle temporal gyrus and discuss the double dis­ sociation of on- and kun-reading. Case report During hospitalization for treatment of a fever of unknown origin, a right-handed, 66-year-old woman with 14 years of education became aware that she could not recall her name. Neurological examination showed right lower homonymous quadrantanopsia, mild anomia, and alexia with agraphia. A brain computed tomo­ graphy (CT) scan revealed subcortical hemorrhage in the left pos­ terior middle temporal gyrus and part of the angular gyrus. Systemic workup revealed ovarian carcinoma with metastatic peri­ toneal dissemination. Magnetic resonance imaging (MRI) per­ formed 30 days after disease onset revealed a high-intensity area in the posterior half of the left middle temporal gyrus that extended to the angular gyrus on T1-weighted and fluid-attenuated inversion recovery (FLAIR) images (Figure 1). 221 Neuropsychological assessment The patient underwent the following neuropsychological assessments from disease onset to 12 months later. Written informed consent was obtained from the patient. Standard neuropsychological tests In the Wechsler Adult Intelligence Scale-III (WAIS-III) adminis­ tered 7 months after disease onset, her general intelligence was within the normal range (Table 1). In the Western Aphasia Battery (WAB, Japanese edition) 3 months after disease onset, minimal anomia and agraphia of kanji were noted. Although with some verbal paralexia, the reading comprehension of mixed kanji and kana sentences was intact. In the Test of Lexical Processing in Aphasia (TLPA) (I. Fujita et al., 2001) 4 months after disease onset, lexical decision (to detect whether a two-character kanji sequence was a real word or not) was nearly normal (93% [149/160] correct). Special neuropsychological tests To determine the characteristics of alexia and agraphia, we administered three kinds of tests during the follow-up period: reading and writing test of single-character kanji and kana transcriptions, on–kun reading test, and multiple character Figure 1. Magnetic resonance images 30 days after disease onset. T1-weighted axial (upper) and sagittal (lower) images and fluid-attenuated inversion recovery (FLAIR) coronal images (middle) indicate a subcortical high-intensity area in the left posterior half of the middle temporal gyrus. 222 M. YOSHIDA ET AL. Table 1. Patient’s scores on standard neuropsychological tests. WAIS-III (7 months after disease onset) Verbal IQ Performance IQ Working memory Processing speed WAB (3 months after disease onset) Spontaneous speech Information content (/10) Fluency (/10) Naming total (/10) Object naming (/60) Repetition (/10) Comprehension total (/10) Reading total (/10) Comprehension of sentences (/40) Kanji printed word–object matching (/3) Kana printed word–object matching (/3) Writing total (/10) Kanji writing (/6) Kana writing (/6) Calculation (/24) Praxis (/60) TLPA (4 months after disease onset) Lexical decision of two-character kanji High-frequency, high-imageable words (/20) High-frequency, low-imageable words (/20) Low-frequency, high-imageable words (/20) Low-frequency, low-imageable words (/20) Nonwords (/80) Total (/160) Patient Controls a Mean (SD) 90 117 90 92 100 (15) 100 (15) 100 (15) 100 (15) 9 10 8.2 54 8.4 9.6 8.7 40 3 3 9.7 2.5 6 24 58 9.7 (0.6) 10.0 (0.0) 9.5 (0.6) 59.2 (2.4) 9.87 (0.3) 9.8 (0.1) 9.5 (0.8) 37.4 (5.3) 3.0 (0.0) 3.0 (0.0) 9.6 (1.0) 5.3 (1.5) 5.7 (1.0) 23.1 (2.3) 59.8 (0.7) 20 19 19 15 76 149 20.00 (0.00) 19.92 (0.35) 19.96 (0.20) 18.83 (2.08) 77.31 (3.76) 156.10 (6.97) a Data for the normal controls are based on the manuals for WASI-III (Japanese edition) (K. Fujita et al., 2006), WAB (Japanese edition) (Sugishita, 1986), and TLPA (I. Fujita et al., 2001). WAIS-III, Wechsler Adult Intelligence Scale-III; WAB, Western Aphasia Battery; TLPA, Test of Lexical Processing in Aphasia. kana reading tests (Table 2). The patient’s responses and time for reading or writing were recorded. Reading and writing test of 100 single-character kanji and kana transcriptions (Table 2, I) All of the kanji characters used in the test are taught in the first 3 years of primary school in Japan (Y. Sakurai et al., 2000b). Kana reading and writing was perfect, whereas kanji writing was severely impaired at 3 months after disease onset. Most of the kanji writing errors were no responses due to a character recall deficit. Although the score for kanji reading was within the normal range, the patient spent much time in reading kanji. Of note is that the patient did not recall on-reading of some kanji characters, even though she could recall the kun-reading. In the second follow-up examination, she was asked to read kanji characters with on-reading only or kun-reading only for a total of 98 kanji characters (two characters that are read solely with on-reading were excluded from the analysis). The result showed that on-reading was selectively impaired compared with kun-reading (P< 0.0001 using Fisher’s exact method). All the characters that were not read with on-reading were read with kun-reading. Occasionally, she read the character first with kun-reading and then searched for the on-reading associated with the kun-reading. Also, she tried to associate two-character on–on reading words that contained the character, and then recalled on-reading (e.g., 白 [shiroi], kun-reading, white → 紅白 [kou-haku], on–on reading, red and white → 白 [haku], onreading). Next, to investigate the effects of visual complexity (mea­ sured by the number of writing stroke sequences), concrete­ ness (Kitao et al., 1977), familiarity (Kitao et al., 1977), frequency (Amano & Kondo, 2000), and imageability (Sakuma et al., 2008), we divided the test characters into two groups (above or under the median)–a more complex, concrete, familiar, or frequent group and a less complex, concrete, familiar, or frequent group–and compared the number of correct responses between the two groups in the follow-up test 5 months after disease onset. In the kanji on-reading only task, complexity (P = 0.04), frequency (P = 0.001), and familiarity (P = 0.006) were significant using Fisher’s exact method. In the kanji writ­ ing task, complexity (P < 0.0001), frequency (P = 0.0048), and familiarity (P < 0.0001) were significant using Fisher’s exact method. These results suggested that visual complexity, fre­ quency, and familiarity influenced the performance of kanji onreading and writing. On–kun reading test of on–on reading, kun–kun reading, and specific- (Jukujikun) reading of two-character kanji words (Table 2, II) The tests consisted of 100 on–on reading words (e.g., 海岸 [kaigan], seashore), kun–kun reading words (e.g., 野原 [no-hara], field), and 60 specific-reading words called Jukujikun (see the Introduction for an example). All of the kanji characters tested here are taught within nine grade levels of compulsory educa­ tion in Japan. The results showed that on-reading was signifi­ cantly more impaired than kun-reading (P < 0.0001 using Fisher’s exact method) and specific-reading (P = 0.003 using Fisher’s exact method), and there was no difference in the percentage of correct scores between kun-reading and specific-reading. The frequency and familiarity (Amano & Kondo, 2000) of these word groups were different (mean logarithmic fre­ quency [SD]: on-reading, 3.47 [0.84]; kun-reading, 2.64 [0.87]; specific-reading, 2.98 [0.74]; mean familiarity [SD] assessed by a seven-rating scale: on-reading, 5.94 [0.43]; kun-reading, 5.62 [0.59]; specific-reading, 5.78 [0.49]). We therefore reduced the number of stimulus items of the on-reading and kun-reading words from 100 to 60 so that the mean and SD would be close to those of the specific-reading (mean logarithmic frequency [SD]: on-reading, 2.97 [0.69]; kun-reading, 2.95 [0.75]; mean familiarity [SD]: on-reading, 5.85 [0.49]; kun-reading, 5.72 [0.50]). In this condition, as well, the inferiority of on-reading to kun-reading or specific-reading remained consistent, and the score for kun-reading was not different from that for specific-reading. Errors included no response (22 out of a total of 42 errors [52%]) and phonological errors (phonemic paralexia, 9/42 [21%]) in the on-reading test, no response (6 out of a total of 15 errors) and phonological errors (6/15) in the kun-reading test, and kun-substitution (changing the specific-reading to kun-reading) (5 out of a total of 11 errors) in the specificreading test. Reading test of kana single characters, three-character kana words and nonwords, and five-character kana words and nonwords (Table 2, III) (Y. Sakurai et al., 2010) Single-character kana were 46 letters of the kana syllabary. Highfamiliarity three-character kana words (mean, 5.98 on a seven- NEUROCASE 223 Table 2. Patient’s scores on special neuropsychological tests. 3 months 5 months I. Reading and writing (100 single-character kanji and kana transcription), raw score and time (percentile) Kanji reading (/100) 97 (2) 96 (1) 4 min 50 s (> 99) Kanji on-reading (/98) n.d. 70 (71.4%) 23 min 9 s Kanji kun-reading (/98) n.d. 96 (98%) 4 min 56 s Kana reading (/100) 100 (99) n.d. Kanji writing (/100) 29 (< 1) Kana writing (/100) 100 (> 39) II. On-kun reading test, raw score and time (percentile) On-reading (/100) On-reading (/60) Kun-reading (/100) Kun-reading (/60) Specific-reading (/60) 48 (< 1) 24 min 29 s (> 99) n.d. 58 (< 1) 9 min 46 s (> 99) 31 (52%) (< 1) 85 (< 1) 12 min 20 s (> 99) 52 (87%) (< 1) 49 (82%) (< 1) 8 min (> 99) 14 months Controls a Mean score (SD) Mean time (SD) 99 (8) 3 min 20 s (> 99) 78 (79.5%) 17 min 25 s 98 (100%) 4 min 5 s 100 (99) 1 min 53 s (90) 50 (< 1) 19 min 12 s (> 99) 100 (> 39) 5 min 52 s (4) 99.6 (0.5) 1 min 13 s (24 s) 95.9 (3.0) 10 min 11 s (136 s) 99.3 (0.9) 8 min 7 s (107 s) 73 (< 1) 10 min 27 s (> 99) 40 (67%) (< 1) 92 (1) 8 min 25 s (> 99) 56 (93%) (10) 45 (75%) (< 1) 6 min 53 s (> 99) 98.6 (1.7) 1 min 26 s (26 s) 58.9 (1.3) 97.3 (2.5) 1 min 58 s (40 s) 58.5 (1.7) 57.6 (3.2) 1 min 17 s (30 s) III. Reading (single-, three-, and five-character kana words and nonwords), correct responses and time (percentile) Single-character kana (/46) 46 (> 19) 44 sb (> 99) Three-character kana words (hf) (/100) 96 (5) 3 min 8 sb (> 99) Three-character kana words (lf) (/100) 99 (67–80) 5 min 26 s (> 99) Three-character kana nonwords (/100) 97 (10–50) 5 min 26 s (> 99) Five-character kana words (/50) 50 (> 49) 2 min 32 sb (> 99) Five-character kana nonwords (/50) 47 (> 39) 4 min 25 s (> 99) 99.6 (1.2) 1 min 32 s (32 s) n.d. n.d. 45.7 (0.6) 24 s (4 s) 98.3 (1.6) 2 min 1 s (37 s) 96.6 (2.0) 2 min 28 s (41 s) 97.8 (1.3) 2 min 1 s (37 s) 49.5 (0.5) 1 min 24 s (32 s) 47.0 (2.1) 2 min 16 s (33 s) a Healthy controls: reading and writing (I), n = 11 (mean, 68 years old); on–kun reading (II), n = 10 (mean, 71 years old); reading (III), n = 11 (mean, 73 years old); all with 9 or more years of education and no past history of neurologic illness. b The mean time for reading kana by the patient (and healthy controls) was as follows: single-character, 0.96 (0.53) s/character; three-character words, 1.88 (1.22) s/ word; and five-character words, 3.04 (1.70) s/word. hf, high familiarity; lf, low familiarity; n.d., not done. point rating scale) and five-character kana words were selected from research reports (Amano & Kondo, 2000; T. Fujita et al., 1991), and three-character and five-character nonwords were made by changing the sequential order of kana characters, e.g., きのう ([kinou], yesterday) → うのき ([unoki], nonword). Lowfamiliarity kana words (mean, 4.20 on a seven-point rating scale) (Amano & Kondo, 2000) were also provided. The results showed that kana reading was nearly normal, and there were no lexicality or familiarity effects. However, the patient read three-character low-familiarity words and nonwords and five-character non­ words in a letter-by-letter manner, and a minor degree of word length effect was observed (Table 2, footnote b). Recovery of function over time Although on-reading alexia and lexical agraphia for kanji recov­ ered partially over the course of the disease, the percentage of correct responses remained at 80% in kanji on-reading and 50% in kanji writing 14 months after disease onset. Discussion The patient presented with alexia with agraphia for kanji, anomia, and acalculia after cerebral hemorrhage in the poster­ ior half of the left middle temporal gyrus extending to the angular gyrus. Although kana words were read correctly, letterby-letter reading and a minor degree of word length effect were observed. Characteristically, alexia of kanji was more pro­ nounced in on-reading: she read some kanji characters with kun-reading, but could not recall on-reading. Agraphia of kanji was due to a character recall deficit. It is clear from the results of two reading tests that onreading was selectively impaired even though kun-reading was relatively preserved. In the 100 single-character kanji read­ ing test, the patient could not recall on-reading of 27 kanji characters that she could read with kun-reading. She was aware that she was able to recall on-reading when she first read the kanji with kun-reading or found an associated twocharacter on–on reading kanji word that contained the target character. The on–kun reading test further demonstrated that 224 M. YOSHIDA ET AL. she was poor at reading on-reading words, whereas she read kun-reading and specific-reading words fairly well. This result, together with our previous finding that kunreading is preferentially disturbed in semantic dementia (Y. Sakurai et al., 2006a), supports a double dissociation between on-reading and kun-reading. Mechanism of on-reading alexia From a neuropsychological perspective (Beeson & Rapcsak, 2010), reading is processed through two pathways: the sublex­ ical phoneme-to-grapheme conversion route and the lexical whole-word route. On the basis of the previous case studies (Y. Sakurai et al., 2010, 2000b; Y. Sakurai, Terao et al., 2008; Y. Sakurai et al., 2006a, 2006b; Sasanuma, 1975) (Figure 2(a)) and our positron emission tomography studies, we proposed an anatomically constrained model of reading (Sakurai, 2004; Y. Sakurai, Momose, et al., 2000a) (Figure 2(b)). In this model, visual information of words and letters is transmitted either to the posterior superior temporal gyrus, the site of the phonolo­ gical lexicon (P), via the lateral occipital gyri and the deep perisylvian temporoparietal cortex (a phonological route that enables successive grapheme–phoneme conversion), or to the posterior inferior temporal cortex (more precisely, mid-fusiform /posterior inferior temporal gyri), the site of the orthographic lexicon (O) or so-called visual word form area (Cohen et al., 2000), via the fusiform gyrus (an orthographic route that enables holistic whole-word recognition). Kana can be pro­ cessed by both routes, whereas kanji are processed exclusively by the orthographic route. Both the posterior superior temporal gyrus (P) and the posterior inferior temporal cortex (O) have a role as the inter­ face to semantics (S). The posterior superior temporal gyrus and the posterior inferior temporal cortex have direct reciprocal connections or indirect interconnections via semantics. It is possible that kanji are recognized and understood along the orthographic-to-semantic route (V → O → S in Figure 2(b)) without reference to its pronunciation (O → P route). As a result, despite having difficulty recalling the on-reading, the patient achieved a full score in reading comprehension of mixed kanji and kana sentences in the WAB test. According to this model, the patient’s on-reading alexia is explained by damage to the connection from the orthographic to the phonological lexicon (O → P in Figure 2(b)). Her lexical decision in the TLPA was within the normal range. Moreover, kun-reading was possible even for those words that could not be read with on-reading. These facts indicate that the ortho­ graphy of words was preserved. In addition, the phonological lexicon was intact because auditory comprehension was nor­ mal. Thus, it is conceivable that access from the visual word form to the corresponding phonetic value was interrupted. On the other hand, the fact that kun-reading and specificreading were relatively preserved can be explained by bypass­ ing the orthography-to-phonology route through semantics. That is, access to the kun-reading becomes possible when the orthography of a word connects with semantics, and then the semantic lexical information connects with the phonological lexical information (O → S→ P in Figure 2(b)). Anatomical substrate of on-reading impairment The lesion was in the posterior middle temporal gyrus. Damage to this area is known to cause agraphia for kanji (Y. Sakurai, Mimura et al., 2008) or lexical agraphia (Ullrich & Roeltgen, 2011) in Western countries. The reason that on-reading alexia accompanied agraphia was that the lesion extended over the posterior half of the middle temporal gyrus, which may have extensively interrupted U-fiber connections from the inferior temporal gyrus to the middle and superior temporal gyri. Alexia with agraphia for kanji and anomia was reported following hemorrhage in the left middle third of the middle temporal gyrus (Yamawaki et al., 2005). The patient did not recall the reading of kanji words, although he could understand the meaning. The study did not report whether on-reading or kun-reading was more impaired. Taking account of the fact that the lesion was situated more anteriorly in the middle temporal gyrus, and thus that semantic memory was more involved, both on-reading and kun-reading may have been impaired to the same degree, or kun-reading and specific-reading may have been more impaired. Our patient showed that on-reading was selectively impaired whereas kun-reading was relatively preserved. Altogether with the aforementioned case, this suggests that the phonetic value of on- and kun-reading is differently represented in the phono­ logical lexicon site; phonetic information of on-reading is located more posteriorly and that of kun-reading is located more ante­ riorly in the superior temporal gyrus. Lexical agraphia for kanji Agraphia was due to a character recall deficit. The mechanism is the same as agraphia due to a middle temporal gyrus lesion (Y. Sakurai, Mimura et al., 2008). We think that access from the phonological lexicon to the orthographic lexicon (P → O in Figure 2(b)) was interrupted or the output of visual word form images from the orthographic lexicon was disrupted. Extension to European languages In summary, the patient showed alexia with agraphia for kanji characterized by on-reading alexia and lexical agraphia. The lesion was in the white matter under the posterior half of the middle temporal gyrus and interrupted fiber connections between the posterior superior temporal gyrus (the site of the phonological lexicon) and the posterior inferior temporal gyrus (the site of the orthographic lexicon), resulting in on-reading alexia and lexical agraphia. In contrast, kun-reading and Jukujikun-reading may have bypassed the orthography-tophonology route through semantics (orthography-tosemantics, and then semantics-to-phonology). The on–kun dissociation supports the view that there are two distinct pathways from the orthographic to the phonological lex­ icon: the direct pathway and the indirect pathway via semantics (Beeson & Rapcsak, 2010). The present study and our previous study (Y. Sakurai et al., 2006a) demonstrate that the two pathways are dissociable. On-reading depends on the direct pathway, whereas kun-reading, which is more associated with semantics, depends on the indirect pathway. If the direct pathway is not NEUROCASE 225 Figure 2. Lesion localization of kanji alexia and kana alexia and the anatomically constrained dual-route reading model. (a) Neural substrates of the postrolandic alexia with regard to kanji and kana. Data are derived from our previous studies (Y. Sakurai et al., 2010, 2000b; Y. Sakurai, Terao et al., 2008; Y. Sakurai et al., 2006a, 2006b) and that of another researcher (Sasanuma, 1975). Kana > Kanji: Kana are more prominent than kanji, post sup T: posterior superior temporal gyrus, lat O: lateral occipital gyri, mid/inf O: middle and inferior occipital gyri, post-Fu/inf O: posterior fusiform and inferior occipital gyri, mid-Fu/inf T: mid-fusiform and inferior temporal gyri, post inf T: posterior inferior temporal cortex, ant T: anterior temporal lobe. (b) Dural-route model of reading. In the phonological route (a), which performs successive grapheme–phoneme conversion, visual information of words and letters goes from the primary visual cortex (V) to the posterior superior temporal gyrus (Wernicke’s area), where the phonological lexicon of words (P) exists, via the lateral occipital gyri and the deep perisylvian temporoparietal cortex. In the orthographic route (b), which performs holistic whole-word recognition, visual information of words and letters goes from the primary visual cortex to the posterior inferior temporal cortex (the so-called visual word form area), where the orthographic lexicon of words (O) exists, via the fusiform gyrus. Kana can be processed by both the orthographic and the phonological routes, whereas kanji are processed exclusively by the orthographic route. Both the posterior superior temporal gyrus and the posterior inferior temporal cortex have a role as the interface to semantics (S). The posterior superior temporal gyrus (P) and the posterior inferior temporal cortex (O) have direct reciprocal connections or indirect interconnections via semantics (S). The patient’s lesion is shown by a gray ellipse, and damaged fiber connections for on-reading alexia are denoted using “X.” Revised with permission from Sakurai (2004). a: Phonological pathway, b: orthographic pathway, O: orthographic lexicon, P: phonological lexicon, S: semantic storage site, V: primary visual cortex. available, only the indirect pathway will be used, making kunreading possible, as in the present patient. In contrast, if semantic storage in the indirect route is damaged, only the direct pathway will be used, enabling on-reading without understanding, as in patients with Japanese semantic dementia. This on–kun dissociation is clinically important because it provides a strategy for rehabilitation of on-reading impairment: semantic association or semantic lexical help by the bypassing (indirect) route may serve to compensate for the impairment. The same phenomenon, if it occurs in European languages, would appear as a symptom in which pronunciation of regular words is recalled more poorly than that of irregular or excep­ tional words that are directly associated with semantics. Similar symptoms may also occur in patients with developmen­ tal dyslexia with abnormal gyrification in the left superior temporal sulcus and middle temporal gyrus (Caverzasi et al., 2018) or changes in white matter in the left temporoparietal region (Peterson & Pennington, 2015). In either case, the semantic association strategy will facilitate the learning of words. Further study is needed to determine whether the same lesion produces impairment of pronunciation recall in European languages. Disclosure statement The authors report no conflicts of interest. ORCID Toshihiro Hayashi http://orcid.org/0000-0003-4362-7439 Yasuhisa Sakurai http://orcid.org/0000-0002-6992-6747 References Amano, S., & Kondo, T. (2000). NTT data base series. Lexical properties of Japanese. First period. Sanseido. Beeson, P. M., & Rapcsak, S. Z. (2010). Neuropsychological assessment and rehabilitation of writing disorders. In J. M. Gurd, U. Kischka, & J. C. Marshall (Eds.), The handbook of clinical neuropsychology (2nd ed., pp. 323–348). Oxford University Press. Caverzasi, E., Mandelli, M. L., Hoeft, F., Watson, C., Meyer, M., Allen, I. E., Papinutto, N., Wang, C., Gandini Wheeler-Kingshott, C. A. M., Marco, E. J., Mukherjee, P., Miller, Z. A., Miller, B. L., Hendren, R., Shapiro, K. A., & GornoTempini, M. L. (2018). Abnormal age-related cortical folding and neurite morphology in children with developmental dyslexia. NeuroImage: Clinical, 18, 814–821. https://doi.org/10.1016/j.nicl.2018.03.012 Cohen, L., Dehaene, S., Naccache, L., Lehericy, S., Dehaene-Lambertz, G., Henaff, M.-A., & Michel, F. (2000). The visual word form area: Spatial and temporal characterization of an initial stage of reading in normal sub­ jects and posterior split-brain patients. Brain, 123(2), 291–307. https:// doi.org/10.1093/brain/123.2.291 Fujita, I., Monoi, H., Okudaira, N., Ueda, M., Ono, K., Furuya, F., Shimogaki, Y., Iguchi, Y., & Sasanuma, S. (2001). A Test of Lexical Processing in Aphasia (TLPA). Assessment of information processing of words (2nd ed.). Escor. Fujita, K., Maekawa, H., Dairoku, H., & Yamanaka, K. (2006). Wechsler Adult Intelligence Scale-third edition (Japanese edition). Nihon Bunka Kagakusha. Fujita, T., Saito, S., & Takahashi, M. (1991). The familiarity values of Japanese five-letter nouns. Memories of Kyoto Tachibana Women’s University, 18, 79–93. Iwata, M. (1986). Neural mechanism of reading and writing in the Japanese language. Functional Neurology, 1(1), 43–52. Kawamura, M. (1990). Localization and symptomatology of pure alexia, pure agraphia and alexia with agraphia. Shinkei Shinrigaku, 6(1), 16–23. http://www.neuropsychology.gr.jp/journal/archive/1990_06_01_03.pdf Kess, J. F., & Miyamoto, T. (1999). The Japanese mental lexicon: Psyxholinguistic studies of kana and kanji processing. John Benjamins. 226 M. YOSHIDA ET AL. Kindaichi, H. (1988). The Japanese language (2nd ed.). Iwanami Shoten. Kitao, N., Hatta, T., Ishida, M., Babazono, Y., & Kondo, Y. (1977). Concreteness, hieroglyphicity and familiarity of kanji (Japanese form of Chinese characters). Shinrigaku Kenkyu: The Japanese Journal of Psychology, 48(2), 105–111. https://doi.org/10.4992/jjpsy.48.105 Miyake, H., & Tanaka, T. (1992). A case of alexia in kana and agraphia in kanji with the left temporo-parietal subcortical hematoma. Shitsugosho Kenkyu, 12(1), 1–9. https://doi.org/10.2496/apr.12.1 Mochizuki, H., & Ohtomo, R. (1988). Pure alexia in Japanese and agraphia without alexia in kanji. The ability dissociation between reading and writing in kanji vs kana. Archives of Neurology, 45(10), 1157–1159. https:// doi.org/10.1001/archneur.1988.00520340111020 Peterson, R. L., & Pennington, B. F. (2015). Developmental dyslexia. Annual Review of Clinical Psychology, 11, 283–307. https://doi.org/10.1146/ annurev-clinpsy-032814-112842 Sakuma, N., Ijuin, M., Fushimi, T., Tatsumi, I., Tanaka, M., Amano, S., & Kondo, T. (2008). NTT data base series. Lexical properties of Japanese. Third period. Sanseido. Sakurai, Y. (2004). Varieties of alexia from fusiform, posterior inferior tem­ poral and posterior occipital gyrus lesions. Behavioural Neurology, 15(3), 35–50. https://doi.org/10.1155/2004/305194 Sakurai, Y. (2011). Lesion localization of non-aphasic alexia and agraphia. Rinsho Shinkeigaku = Clinical Neurology, 51(8), 567–575. https://doi.org/ 10.5692/clinicalneurol.51.567 Sakurai, Y., Asami, M., & Mannen, T. (2010). Alexia and agraphia with lesions of the angular and supramarginal gyri: Evidence for the disruption of sequential processing. Journal of the Neurological Sciences, 288(1–2), 25–33. https://doi.org/10.1016/j.jns.2009.10.015 Sakurai, Y., Mimura, I., & Mannen, T. (2008). Agraphia for kanji resulting from a left posterior middle temporal gyrus lesion. Behavioural Neurology, 19 (3), 93–106. https://doi.org/10.1155/2008/393912 Sakurai, Y., Momose, T., Iwata, M., Sudo, Y., Ohtomo, K., & Kanazawa, I. (2000a). Different cortical activity in reading of Kanji words, Kana words and Kana nonwords. Brain Research. Cognitive Brain Research, 9 (1), 111–115. https://doi.org/10.1016/S0926-6410(99)00052-X Sakurai, Y., Takeuchi, S., Takada, T., Horiuchi, E., Nakase, H., & Sakuta, M. (2000b). Alexia caused by a fusiform or posterior inferior temporal lesion. Journal of the Neurological Sciences, 178(1), 42–51. https://doi.org/10. 1016/S0022-510X(00)00363-4 Sakurai, Y., Terao, Y., Ichikawa, Y., Ohtsu, H., Momose, T., Tsuji, S., & Mannen, T. (2008). Pure alexia for kana. Characterization of alexia with lesions of the inferior occipital cortex. Journal of the Neurological Sciences, 268(1–2), 48–59. https://doi.org/10.1016/j.jns.2007.10.030 Sakurai, Y., Tsuchiya, K., Oda, T., Hori, K., Tominaga, I., Akiyama, H., Bando, M., Haga, C., Iwata, M., & Mannen, T. (2006a). Ubiquitin-positive frontotemporal lobar degeneration presenting with progressive Gogi (word-meaning) apha­ sia. A neuropsychological, radiological and pathological evaluation of a Japanese semantic dementia patient. Journal of the Neurological Sciences, 250(1–2), 3–9. https://doi.org/10.1016/j.jns.2006.05.067 Sakurai, Y., Yagishita, A., Goto, Y., Ohtsu, H., & Mannen, T. (2006b). Fusiform type alexia: Pure alexia for words in contrast to posterior occipital type pure alexia for letters. Journal of the Neurological Sciences, 247(1), 81–92. https://doi.org/10.1016/j.jns.2006.03.019 Sasanuma, S. (1975). Kana and Kanji processing in Japanese aphasics. Brain and Language, 2(3), 369–383. https://doi.org/10.1016/S0093-934X(75) 80077-0 Sugishita, M. (1986). The Western Aphasia Battery, Japanese edition. IgakuShoin. Ullrich, L., & Roeltgen, D. P. (2011). Agraphia. In K. M. Heilman & E. Valenstein (Eds.), Clinical neuropsychology (5th ed., pp. 130–151). Oxford University Press. Yamawaki, R., Suzuki, K., Tanji, K., Fujii, T., Endo, K., Meguro, K., & Yamadori, A. (2005). Anomic alexia of kanji in a patient with anomic aphasia. Cortex, 41(4), 555–559. https://doi.org/10.1016/S0010-9452(08) 70195-3