Neuropsychologia 53 (2014) 264–273

Contents lists available at ScienceDirect

Neuropsychologia journal homepage: www.elsevier.com/locate/neuropsychologia

Multiple thalamo-cortical disconnections in anterior thalamic infarction: Implications for thalamic mechanisms of memory and language Yoshiyuki Nishio a,n, Mamoru Hashimoto b, Kazunari Ishii c, Daisuke Ito d, Shunji Mugikura d, Shoki Takahashi d, Etsuro Mori a a

Department of Behavioral Neurology and Cognitive Neuroscience, Tohoku University Graduate School of Medicine, 2-1, Seiryo-machi, Aoba-ku, Sendai Japan Department of Psychiatry and Neuropathology, Faculty of Medical and Pharmaceutical Sciences, Kumamoto University, Kumamoto, Japan Department of Radiology, Kinki University School of Medicine, Sayama, Japan d Department of Diagnostic Radiology, Tohoku University Graduate School of Medicine, Sendai, Japan b c

art ic l e i nf o

a b s t r a c t

Article history: Received 16 August 2013 Received in revised form 27 November 2013 Accepted 29 November 2013 Available online 7 December 2013

Amnesia and linguistic deficits that are associated with thalamic damage have attracted the attention of researchers interested in identifying the neural networks involved in memory and language. The Papez circuit, which is composed of the hippocampus, mammillary body and anterior thalamic nuclei, was first proposed to be critical for memory. However, subsequently, the roles of the neural circuit consisting of the rhinal/parahippocampal cortices and the mediodorsal thalamic nuclei became evident. The ventral lateral nuclei or its adjacent structures have been found to be involved in semantic processing, but the specific neural circuits dedicated to language functions have not been identified. Anterior thalamic infarcts, which affect very circumscribed regions of the ventral anterior portion of the thalamus, often cause paradoxically prominent memory and language deficits. We conducted tractography analyses in 6 patients with left anterior thalamic infarcts to identify neural connections or circuits in which disruptions are associated with memory and language deficits in this condition. The current study demonstrated that the mammillothalamic tract, which connects the mammillary body with the anterior thalamic nuclei, and the anterior and inferior thalamic peduncles, which contain neural fibers that extend from several thalamic nuclei to the anterior temporal, medial temporal and frontal cortices, are disrupted in anterior thalamic infarction. These extensive thalamo-cortical disconnections appear to be due to the dissection of the neural fibers that penetrate the ventral anterior nucleus of the thalamus. Our results suggest the following: (1) amnesia that is associated with anterior thalamic infarction is best interpreted in the context of dual/multiple-system theories of memory/amnesia that posit that multiple neural circuits connecting the anterior and mediodorsal thalamic nuclei with the hippocampus and rhinal/parahippocampal cortices work in concert to support memory function; and (2) the semantic deficits observed in this syndrome may be associated with thalamo-anterior temporal and thalamo-lateral frontal disconnections. & 2013 Elsevier Ltd. All rights reserved.

Keywords: Language Medial temporal lobe Memory Semantic Thalamus

1. Introduction 1.1. A historical overview of diencephalic amnesia The role of thalamic damage in the amnesia observed in alcoholics (i.e., Wernicke–Korsakoff syndrome) was uncovered in the 1930s (Aggleton, 2008; Bender & Schilder, 1933; Kopelman, 1995). This modest discovery preceded Scoville and Milner's

Abbreviations: AN, anterior thalamic nuclei; IML, internal medullary lamina; MD, mediodorsal nuclei; MTT, mammillothalamic tract; VA, ventral anterior nucleus; VAmc, magnocellular ventral anterior nucleus; Vim, ventral intermediate nucleus; VLa, ventral lateral anterior nucleus; VLp, ventral lateral posterior nucleus n Corresponding author. Tel.: þ 81 22 717 7358; fax: þ81 22 717 7360. E-mail address: [email protected] (Y. Nishio). 0028-3932/$ - see front matter & 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.neuropsychologia.2013.11.025

landmark discovery of medial temporal amnesia by 20 years (Scoville & Milner, 2000). Subsequent studies demonstrated striking similarities between the neuropsychological features of thalamic and medial temporal amnesias and led to the idea that memory functions are supported by a network that consists of these neural structures. Various theories have been proposed to synthesize the empirical findings on medial temporal lobe mechanisms of memory (Eichenbaum, Yonelinas, & Ranganath, 2007; Saksida & Bussey, 2010; Wixted & Squire, 2011), whereas fewer attempts have been made to do the same for the counterpart of this structure; i.e., the thalamus (Aggleton & Brown, 2006). Empirical evidence concerning thalamic amnesia has originated from several different lines of research that include studies of Wernicke–Korsakoff syndrome, thalamic infarcts and animal

Y. Nishio et al. / Neuropsychologia 53 (2014) 264–273

neuropsychological studies. Integration of the findings from these research fields is essential for further understanding of the neural mechanisms of memory. Although a series of neuropathological studies have revealed that the mammillary body, anterior thalamic nuclei (AN) and mediodorsal nuclei (MD) are the primary loci of neurodegeneration in Wernicke–Korsakoff syndrome (Kopelman, 1995; Malamud & Skillicorn, 1956), the roles of each of these diencephalic structures in memory functions have been matters of debate. In a study that is probably one of the most systematic investigations currently available, Harding, Halliday, Caine, and Kril (2000) demonstrated that significant degeneration of the AN is found only in patients with Wernicke–Korsakoff syndrome and not in patients with Wernicke's encephalopathy in the absence of profound amnesia. On the basis of these findings, the role of the AN in memory has been emphasized (Aggleton, 2008; Harding et al., 2000). However, the study by Harding and colleagues demonstrated that the mammillary body and MD were consistently affected in Wernicke's encephalopathy and Wernicke–Korsakoff syndrome irrespective of the presence or absence of profound amnesia. These findings allow alternative interpretations in which simultaneous damage to these structures or the amount and severity of diencephalic pathology contribute to the development of profound amnesia. One of the earliest, but still influential, hypotheses concerning the neural network of memory was proposed by Delay and Brion (Gaffan & Wilson, 2008; Mayes, 2000). Their theory assigns a central role to the Papez circuit, which arises from the hippocampus, travels through the fornix, mammillary body, mammillothalamic tract (MTT), AN and posterior cingulate cortex, and then returns to the hippocampus. Although the importance of the AN and hippocampus is stressed in this theory, subsequent neuropsychological studies in monkeys have demonstrated that neither isolated damage to the AN nor damage to the hippocampus results in dense amnesia, and additional lesions in the MD or the rhinal cortices are necessary for the development of severe memory impairments (Gaffan & Parker, 2000; Gaffan, Parker, & Easton, 2001; Mishkin, 1978; Murray & Mishkin, 1998; Parker & Gaffan, 1997). These discoveries promoted the development of new theories that embrace both the contributions of structures outside the Papez circuit and the neural network views of memory; these dual- or multi-system theories propose that distinct diencephalic and medial temporal structures form parallel, but partially interconnected, neural circuits (i.e., the hippocampalAN (Papez) and rhinal-MD circuits), and that these two neural circuits work in concert to support memory functions (Aggleton & Brown, 1999, 2006; Mishkin, 1982). Specifically, some authors hypothesize that the hippocampal-AN circuit supports the process of recollection, which may be involved in the retrieval of detailed information associated with previous experiences, whereas the rhinal-MD circuit may play a pivotal role in the process of familiarity, which is a fundamental process in recognition memory (Aggleton & Brown, 1999; Yonelinas, Aly, Wang, & Koen, 2010). However, there has been considerable controversy regarding both the familiarity/ recollection distinction of memory and their neural substrates (Wixted & Squire, 2011). As the theories for neural networks of memory have been developed, there have been attempts to understand amnesia associated with Wernicke–Korsakoff syndrome in relation to neural network disruption. Previous PET studies consistently demonstrated decreased glucose metabolism in extensive cortical regions, including the frontal, temporal and parietal cortices, in Wernicke–Korsakoff syndrome, suggesting that thalamo-cortical network disruptions are not restricted to the AN-hippocampal circuit but may be more widespread in this disorder (Paller et al., 1997; Reed et al., 2003). Memory research actively focused on the amnesia associated with isolated thalamic infarction with the advent of CT and MRI in the 1980s and 1990s, respectively. Thalamic infarcts are classified

265

into 4 subtypes on the basis of arterial territories: anterior, paramedian, inferolateral and posterior thalamic (Carrera & Bogousslavsky, 2006; Schmahmann, 2003). Among these subtypes, anterior and paramedian thalamic infarcts have received particular attention because their primary clinical manifestations are amnesia and other cognitive deficits. Another remarkable aspect of these thalamic vascular syndromes is that very small and circumscribed lesions often produce profound memory sequelae, which allows researchers to examine the roles of single or limited thalamic structures that are critical for memory. Because the AN, which is considered to be one of the most important diencephalic structures in Wernicke–Korsakoff syndrome and the Delay–Brion hypothesis, is spared in most cases of anterior and paramedian thalamic infarcts, the following 3 hypothetical mechanisms of thalamic infarct-associated amnesia have been proposed: (1) disruption of the Papez circuit due to damage to the MTT; (2) damage to the MD; and (3) disruption of the MDrhinal cortical circuit due to dissection of the neural fibers that pass through the internal medullary lamina (IML) (Graff-Radford, Tranel, Van Hoesen, & Brandt, 1990; Mori, Yamadori, & Mitani, 1986; von Cramon, Hebel, & Schuri, 1985). Most of these studies are based on single or small numbers of cases because isolated thalamic infarcts are relatively rare. Therefore, several authors have attempted to integrate the findings from the relevant case studies in literature reviews. Van der Werf, Witter, Uylings, and Jolles (2000) analyzed the locations of lesions in 60 published cases of isolated thalamic infarcts and concluded that damage to the MTT is crucial for the development of amnesia. Additionally, in a review of 83 cases, Carlesimo, Lombardi, and Caltagirone (2011) demonstrated that 95% of patients with MTT damage and 46% of those without MTT damage had amnesia (chi-square 25.3; po 0.01), whereas the involvement of neither the MD nor the IML (referred to as the ‘intralaminar nuclei’ by these authors) predicts the development of amnesia . In the discussions of these review papers, the roles of the MTT and the Papez circuit was emphasized and less attention was given to the contributions of other thalamic structures and related neural circuits. However, the review by Carlesimo et al. (2011) reported that half of the patients without MTT damage developed amnesia, suggesting mechanisms other than the disruption of the Papez circuit in the thalamic infarct associated-amnesia. 1.2. Theoretical and methodological issues in studies of memory impairment associated with thalamic infarcts Here, we review inconsistencies between the current interpretations of the empirical findings from animal and human studies of thalamic amnesia and between the theories derived from those studies. Neuropsychological studies in animals have demonstrated that disruptions of two or more of the neural circuits that consist of distinct thalamic and medial temporal structures are necessary for profound memory impairment; thus, these studies have led to proposals that these multiple neural circuits work in concert to support memory function (Aggleton & Brown, 1999, 2006; Mishkin, 1982; Zola-Morgan & Squire, 1993). Conversely, the studies of Wernicke–Korsakoff syndrome and thalamic infarcts have been interpreted to indicate that profound amnesia can be ascribed to lesions that are restricted to the AN or the MTT (for a counter view, see Paller et al., 1997). The supposition that lesions to the MTT alone are capable of producing profound amnesia is consistent with the idea that the integrity of the Papez circuit is most critical for the maintenance of normal memory function (Aggleton, 2008; Carlesimo et al., 2011; Harding et al., 2000; Van der Werf et al., 2000). These differences in interpretation may partially originate from the complicated nature of lesions in human patients. In animal studies, targeted thalamic nuclei, such

266

Y. Nishio et al. / Neuropsychologia 53 (2014) 264–273

as the AN and MD, are selectively and completely destroyed. In contrast, the lesions of Wernicke–Korsakoff syndrome are diffuse and vary in severity (Kopelman, Thomson, Guerrini, & Marshall 2009; Pitel et al., 2012). Moreover, in most cases of thalamic infarcts, individual thalamic nuclei are only partially affected, and intrathalamic white matter structures, such as the MTT and IML, are also involved (Carrera & Bogousslavsky, 2006; Kopelman, 1995; Schmahmann, 2003). As previously described, it is impossible to reach clear conclusions regarding which thalamic structures are critical for memory based exclusively on the results of studies of Wernicke–Korsakoff syndrome. The results of previous clinicopathological studies of this disorder can be interpreted not only in the framework of the Delay–Brion theory but also in the frameworks of dual-/multi-system theories. Regarding thalamic infarcts, we argue that the interpretations presented in previous case studies and literature reviews may be profoundly biased by the following confounding factors. First, lesion overlap approaches that examine a limited number of regions of interests may result in neglect of the contributions of structures and circuits outside of those regions (Carlesimo et al., 2011; Van der Werf et al., 2000). This bias is present in all lesion overlap studies but affects studies of thalamic infarcts particularly profoundly because the connectional anatomy of the intrathalamic fiber tracts is largely unknown, and memoryrelated neural circuits can be disrupted at unexpected intrathalamic regions. Second, the neuroanatomical analyses in more than half of the previous studies of thalamic infarcts are descriptive, nominal and do not include detailed neuroimaging analyses. These studies, and the reviews based on them, do not provide any information regarding the subregions of individual thalamic nuclei that are involved (for example, the anterior, posterior, superior or inferior portions of a nucleus) or whether the lesions affected only a small portion of a particular thalamic structure or the whole of that structure. These biases could easily be overcome by neuroimagingbased lesion overlap analyses with regions of interest that include all of the thalamic regions that can be affected by anterior and paramedian thalamic infarcts (Nishio, Hashimoto, Ishii, & Mori, 2011; Pergola et al., 2013). 1.3. The thalamus and language/semantic processing It has been known since the early 20th century that damage to the dominant thalamus occasionally results in aphasia (De Witte et al., 2011). However, information concerning which thalamic nuclei are involved in language processing did not become available because severe and persistent language problems occur only after large thalamic hemorrhages. The relationship between language functions and specific thalamic nuclei was first suggested in the 1960s with the advent of stereotactic surgeries of the thalamus for Parkinson's disease. Notably, a series of experiments by Ojemann and colleagues demonstrated that electrical stimulation to the ventral lateral portion of the dominant thalamus evokes anomia and other linguistic deficits (Ojemann, 1977; Ojemann & Ward, 1971). However, language complications have rarely been observed in the later era of MRI-guided thalamotomy and deep brain stimulation targeting the ventral intermediate nucleus (Vim; ventral lateral posterior nucleus (VLp) in the terminology of Jones) (Jones, 2007; Schmahmann, 2003; Voon, Kubu, Krack, Houeto, & Troster, 2006). The language symptoms observed in earlier electrical stimulation studies were likely due to imprecise anatomical targeting. Indeed, a recent intracerebral EEG study by Wahl et al. (2008) demonstrated that evoked responses associated with semantic processing can be observed in monopolar recordings from Vim electrodes with mastoid references but not by bipolar recordings from electrodes within the Vim . The authors suggested that this evoked potential is generated in thalamic nuclei near the Vim, such as the intralaminar nuclei.

Additional evidence for the roles of specific thalamic nuclei in language has come from studies of isolated thalamic infarcts. A variety of semantic processing deficits have been reported, particularly after anterior thalamic infarcts, including proper name anomia (Semenza, Mondini, & Zettin, 1995), category-specific anomia (Levin, Ben-Hur, Biran, & Wertman, 2005), object use deficits (Rai et al., 2004) and semantic retrieval impairments (Pergola et al., 2013; Segal, Williams, Kraut, & Hart, 2003). These types of semantic deficit are also observed in patients with lesions of the anterior temporal lobe, temporo-occipital region and lateral frontal lobe (Fukatsu, Fujii, Tsukiura, Yamadori, & Otsuki, 1999; Gainotti, 2000; Patterson, Nestor, & Rogers, 2007; Semenza et al., 1995). Moreover, functional neuroimaging studies have demonstrated that these cortical regions are involved in semantic processing in normal subjects (Binder & Desai, 2011; Simmons & Martin, 2009; Thompson-Schill, 2003). These findings lead to the hypothesis that one or more of the structures affected by anterior thalamic infarcts (e.g., the ventral anterior and ventral lateral nuclei) form neural networks with the anterior temporal, temporo-occipital and lateral frontal cortices to support semantic processing. This idea is partially supported by the result of our previous PET study in which remote effects or diaschisis in the anterior temporal and lateral frontal cortices were observed in patients with isolated anterior thalamic infarcts (Nishio et al., 2011). However, the specific thalamic structures that are crucial for semantic processing and other language functions remain to be identified.

1.4. The aim of the current study Neuroanatomical information that identifies the thalamic structures and thalamo-cortical connections that are involved in thalamic infarcts is fundamental for the elucidation of the thalamic mechanisms of memory and language. In this study, we addressed this issue by using MRI lesion overlap analyses with stereotactic neuroanatomical localization and diffusion tensor tractography analyses in a series of patients with anterior thalamic infarcts who developed verbal memory deficits and language problems. The results will be discussed in relation to the gap between interpretations that are based on animal neuropsychological and human patient studies of thalamic amnesia and in relation to the thalamo-cortical networks that support language/semantic processing.

2. Methods 2.1. Subjects We consecutively recruited six right-handed patients (mean age 767 7.4 years; 2 females; mean years of education, 9.2 72.9) in the subacute phase of isolated anterior thalamic infarction from the patients who were admitted to the Hyogo Institute for Aging Brain and Cognitive Disorders, a research-oriented dementia clinic, from 1993 to 2001. The patients were identical to the subjects of our previous study, and their clinical characteristics are described in Tables 1 and 2, Supplementary material and elsewhere (Nishio et al., 2011). Briefly, we included patients who had circumscribed infarcts of the anterior portion of the thalamus and lacked lesions elsewhere based on MRI examinations. All of the patients presented to our hospital with sudden onset of cognitive or behavioral problems, such as forgetfulness, loss of spontaneity and dysnomia. All patients were free of histories of other neurological or psychiatric diseases and had no previous episodes that suggested premorbid cognitive impairment. Lesions were located in the left thalami of all patients (the lesional volumes normalized by total brain size for individual patients are shown in Table 1).The duration between the onset of symptoms and the initiation of examination ranged between 1 and 4 weeks (mean 37 1.3 weeks). All procedures used in this study were approved by the Ethics Committees of the Hyogo Institute and Tohoku University School of Medicine. Written informed consent was obtained from the patients and their relatives and from the control subjects.

Y. Nishio et al. / Neuropsychologia 53 (2014) 264–273

267

Table 1 Demographic and clinical profiles of the patients. Patient no.

1

2

3

4

5

6

Age Sex Handedness Education (years) Medical history Neurological findings Duration (week)

81 M R 8 HT None 3 Dizziness, urinary incontinence, dysnomia, forgetfulness

80 F R 8 HT, RA None 4

75 M R 14 HT, DM None 1

66 M R 8 HT, DM None 4

Dysnomia, forgetfulness

Severe apathy, forgetfulness

85 F R 6 HT None 2 Confusion, apathy, forgetfulness, topographical disorientation

69 M R 11 HT None 4 Visual hallucinations, forgetfulness, executive dysfunction

224

216

350

762

700

Symptoms at onset Normalized lesional volume (ml)

Apathy, dysnomia

456

Abbreviations: HT, hypertension; RA, rheumatoid arthritis; and DM, diabetes mellitus.

Table 2 Results of the neuropsychological tests and behavioral observations. Patient no.

1

2

3

4

5

6

MMSE

Total (30)

25

27

27

16a

24

22a

WAIS-R

VIQ PIQ Information Digit span Vocabulary Arithmetic Comprehension Similarities Picture completion Picture arrangement Block design Object assembly Digit symbol Verbal memory index Visual memory index General memory index Attention/concentration index Delayed recall index

68a 85 6a 6a 5a 7 2a 4a 8 9 5a 9 6a o 50a 72a o 50a 66a o 50a

78a 106 6a 6a 7 8 7 5a 11 12 13 11 8 64a 93 71a 84a o 50a

88 93 10 7 7 12 5a 8 9 6a 9 11 9 61a 118 77a 94 69a

65a 66a 5a 4a 5a 7 2a 4a 5a 5a 4a 3a 4a o 50a 68a o 50a 55a o 50a

81a 91 6a 7 6a 8 7 8 8 7 12 9 7 50a 114 71a 77a 71a

89 91 8 12 9 10 5a 6a 9 6a 11 9 8 64a 100 73a 97 83a

(-) 69.2a 13a 7.2a 8.9a 5.5a 6.7a 6.4a 9.7 9.7 74a 4a 0a 25 ✓   ✓ Apathy

(-) 90.8a 17a 9.8 9.6 9 10 9.7 9.8 9.8 97 10 3a 30   ✓ ✓ Apathy

(-) 86.4a 17a 7.7a 9.9 8.6a 8.9 9.9 9.8 9.7 84 9 1a 23 ✓  ✓ ✓ Apathy

(-) 71a 12a 7.2a 9.2 7.1a 4.1a 4a 7.8a 7.8a 66 4a 1a 14a  ✓ ✓ ✓ Apathy

(-) 83.6a 16a 9a 9.9 6.9a 7.2a 9.1 10 10 86 4a 2a 32 ✓  ✓ ✓ Apathy

(-) 87.6a 16a 9.5a 10 8.3a 7.7a 8.9 10 10 89 12 7 26 ✓ ✓ ✓ ✓ Apathy

WMS-R

Retrograde amnesia WAB AQ Spontaneous speech (20) Auditory comprehension (10) Repetition (10) Naming (10) Reading (10) Writing (10) Praxis R (10) Praxis L (10) Picture Naming (1 0 0) Animal fluency (/min) Initial fluency (/min) RCPM (36) Color-form sorting Fist-edge-palm 2-1 tapping Alternative pattern drawing Behavior

Abbreviations: MMSE, Mini-Mental State Examination; RCPM, Raven's colored progressive matrices; WAIS-R, Wechsler Adult Intelligence Scale-Revised; WMS-R, Wechsler Memory Scale-Revised; WAB, Western Aphasia Battery; VIQ, verbal intelligence quotient; PIQ, performance intelligence quotient; and AQ, aphasia quotient. ✓, passed; and  , failed. The scores on the WAIS-R subtests are indicated with respect to the scaled score. a

Score below the mean - 1SD of the normative data.

2.2. Lesion localization Coronal three-dimensional T1-weighted images (SPGR sequence; TR, 14 ms; TE, 3 ms; flip angle, 201; resolution, 1.5  0.86  0.86 mm) were obtained using a 1.5-T GE Signa Horizon system. The images were reconstructed into 1.0-mm isotropic transverse sections and then normalized to the MNI-T1 template using the affine

transformation algorithm implemented in SPM5 (http://www.fil.ion.ucl.ac.uk/spm/ software/spm5/). In a preliminary analysis, we compared two different methods of spatial normalization, simple affine (linear) transformation and nonlinear transformation with cost function masking, both of which have been proposed for the analysis of brains with focal lesions (Brett, Leff, Rorden, & Ashburner, 2001). The positions of the lesions relative to the third ventricle and the boundaries of the

268

Y. Nishio et al. / Neuropsychologia 53 (2014) 264–273

thalamus were preserved during normalization with simple affine transformations, whereas the lesions were located more laterally than the original lesions after nonlinear normalization with cost function masking. On the basis of these preliminary results, we chose simple affine normalization for the current analysis. We manually traced the patients’ lesions on the normalized images. The obtained lesion masks were used for localization of the lesions and for the tractography analyses described below. The localization of the thalamic and adjacent structures that were involved was stereotactically determined using an electronic version of the Schaltenbrand–Wahren (S–W) atlas (Nowinski, Thirunavuukarasuu, & Benarbid, 2005). Additional details have been described elsewhere (Nishio et al., 2011).

anterior thalamic infarction. Using the ‘winner-take-all’ method of connectivitybased segmentation, each voxel within the lesion seed was classified according to the target ROI that was most densely connected with that voxel (Behrens et al., 2003). The seed segmentation maps obtained from the 10 normal subjects were overlapped and thresholded to display the voxels that were common to 5 or more subjects.

2.3. Tractography

The results are shown in Fig. 1A. The ventral anterior proper (VA proper; also referred to as the provocellular VA or the VA), magnocellular ventral anterior nucleus (VAmc), ventral lateral anterior (VLa), ventral lateral posterior (VLp), reticular (R) nuclei and mammillothalamic tract (MTT) were damaged in all patients. The anterior nuclei (AN) were preserved in all patients. The mediodorsal nuclei (MD), internal medullary lamina (IML)/central medial nucleus and capsular genu were affected in 1, 3 and 3 patients, respectively. These results have been described in more detail elsewhere (Nishio et al., 2011).

We acquired diffusion-weighted images (TR, 8368 ms; TE, 6.6 ms; flip angle, 801; SENSE, 2; thickness, 2 mm; field of view, 224  224 mm; matrix, 112  112 mm; resolution, 2  2  2 mm) and three-dimensional T1-weighted images (MPRAGE sequence; TR, 6.6 ms; TE, 3 ms; flip angle, 81; SENSE, 2; thickness, 1 mm; field of view, 240  240 mm; matrix, 240  240 mm; resolution, 1  1  1 mm) of 10 healthy subjects (31.2 7 8.2 years; 7 men and 3 women) using a 3.0-T Philips Achieva scanner with an 8-channel head coil. Diffusion gradients were applied along 64 directions with a b-value of 1000 s/mm2. Three sets of diffusion-weighted data were averaged to increase the signal-to-noise ratio. Tractrography analyses were performed using the FMRIB Diffusion Toolbox of the FSL Software Library (http://www.fmrib.ox.ac.uk/fsl/). Based on a multi-fiber diffusion model, the probability distributions in the direction of approximately 2 fiber populations were estimated for each brain voxel (Behrens, Berg, Jbabdi, Rushworth, & Woolrich, 2007). We generated 5000 streamline samples from each voxel to construct connectivity distributions for the whole brains. For each voxel, the probabilities of connections with the other voxels were represented as the numbers of streamline samples (per 5000) that passed through that voxel. Wholebrain connectivity distribution maps created in the native space were transformed onto the standard space using an affine transformation algorithm with the MPRAGE image of each subject and the MNI T1 template. To reconstruct the fiber tracts that were disrupted in patients with anterior thalamic infarcts, a “negative fiber-tracking” procedure was used (Gutman, Holtzheimer, Behrens, Johansen-Berg, & Mayberg, 2009; Molko et al., 2002). We conducted 3 different types of analyses: (1) seed-only analysis, (2) seed-and-target analysis and (3) connectivity-based seed segmentation analysis. The seed-only tractography analysis included two steps, normal subject-level analysis and patient-level analysis. First, 10 whole-brain connectivity distribution maps were obtained from 10 normal subjects, which were assigned to each of the 6 patients. We defined a lesion mask individually for each patient (see Section 2.2.) and used it as a seed region of interest (seed ROI). For each patient, tractography was performed on the 10 whole-brain connectivity maps using the patient-specific seed ROI. For each of the 10 resultant tractography maps, only voxels to which 100 or more streamlines reached were displayed (thresholded at Z 100 out of 5000 streamlines) and then transformed into a binary image (Ciccarelli et al., 2006). The obtained binary maps were overlapped and thresholded to display only those voxels that were shared by 5 or more normal subjects. This process produced a mean connectivity map for each patient (normal subject level analysis) (Gutman et al., 2009). We chose this relatively lenient threshold because more stringent thresholds would have displayed, for example, only the voxels that fully overlapped across all 10 subjects and would have included the effects of variables of no interest, such as individual differences and variability in image quality across normal subjects, in the patient-level analysis. A stringent threshold criterion that displayed only the voxels that overlapped across the mean connectivity maps of the 6 patients was used in the subsequent patient-level analyses. The voxels that survived this criterion represent the fiber tracts that are commonly disrupted in the patients with anterior thalamic infarction. In the seed-and-target analysis, we performed identical procedures to those previously described but used target ROIs instead. The aim of this analysis was to delineate more detailed trajectories of the fiber tracts. We chose the lateral frontal lobe (middle frontal gyrus), the orbital frontal lobe, the anterior temporal lobe and the medial temporal lobe as target ROIs because the connections between the thalamus and these brain regions have been demonstrated in animal and human studies (Jones, 2007; Schmahmann, 2003) and because dysfunctions of these brain regions were illustrated in our previous PET study of anterior thalamic infarcts (Nishio et al., 2011). The locations and spatial extents of the four ROIs were determined according to the Harvard-Oxford cortical structural atlas distributed with the FSL package. The connectivity-based seed segmentation analysis included a normal subjectlevel analysis. The sum (union) of the lesions of the 6 patients was used as a lesion seed. The four target ROIs were identical to those used in the seed-and-target analysis. Although a tractography-based connectivity atlas of the thalamus was previously available (Behrens et al., 2003), this atlas did not include connectivity maps between the individual thalamic nuclei and the distinct subdivisions of the prefrontal and temporal cortices. We aimed to obtain more detailed information about the connectional neuroanatomy of the thalamic structures involved in

3. Results 3.1. Lesion localization

3.2. Seed-only analysis A negative fiber-tracking analysis delineated massive thalamocortical connections of the ipsilateral frontal and temporal lobes and fewer connections with the medial parietal lobe, occipital lobe, dorsal upper brainstem and cerebellum (Fig. 2). Close inspection revealed that the mammillothalamic tract was depicted as a disrupted fiber tract (Fig. 3). Some transcallosal fibers were also found. 3.3. Seed-and-target analysis A cartoon of the thalamic nuclei and related fiber bundles is shown in Fig. 4A to help understand the anatomical details described below. The lesion sites were connected with the ipsilateral middle frontal gyrus (the cyan area in the transverse sections shown in Fig. 4B) through the anterior and superior thalamic peduncles (shown in violet in Fig. 4B) and with the orbital frontal lobe (the dark blue area in the transverse sections shown in Fig. 4B) through the inferior thalamic peduncle (shown in green in Fig. 4B). Two distinct fiber tracts to the medial and anterior temporal lobes were delineated as disrupted tracts (shown in cyan and dark blue in the transverse sections of Fig. 4C, respectively): the first tract corresponds to the inferior thalamic peduncle (Fig. 4C), and the second tract corresponds to the temporopulvinar bundle of Arnold (Saunders, Mishkin, & Aggleton, 2005) (also referred to as the ventral subcortical bundle (Schmahmann & Pandya, 2006)), which arose from the anterior portion of the thalamus, passed posteriorly through the pulvinar and then turned anteriorly at the temporal stem to reach the medial and anterior temporal lobes (Fig. 4C). The fiber pathways to the medial temporal lobe and the anterior temporal lobe were largely overlapping, and no clear differences were detected between the trajectories of these pathways. 3.4. Connectivity-based seed segmentation analysis The medial, lateral and intermediate portions of the regions involved in anterior thalamic infarcts demonstrated the following differences in connectivity (Fig. 1B, Table 3): the medial portion, including the VA proper, VAmc, MTT, MD and IML, exhibited the strongest connections with the medial temporal lobe; the lateral portion, including the VA proper, VLa and the dorsal portion of the VLp, exhibited predominant connections with the middle frontal

Y. Nishio et al. / Neuropsychologia 53 (2014) 264–273

269

Fig. 1. Results of the stereotactic lesion localization (A) and the connectivity-based seed segmentation analysis (B). A, Transverse images from the Schaltenbrand-Wahren atlas are shown in the left column. Images showing lesions (red) superimposed on the Montreal Neurological Institute (MNI) template are indicated in the right column. The voxels that overlapped in more patients are colored in brighter red. B, The seed region, which was the sum of the lesions of 6 patients, was classified according to the connectivities to the medial temporal lobe (MTL)/anterior temporal lobe (ATL), middle frontal gyrus (MFG) and orbital frontal lobe (OFL). Abbreviations: VA, ventral anterior nucleus proper; VAmc, magnocellular ventral anterior nucleus; VLa, ventral lateral anterior nucleus; VLp, ventral lateral posterior nucleus; MTT, mammillothalamic tract; and IML, internal medullary lamina.

gyrus; and the intermediate portion, including the VA proper, VAmc, MTT, VLa and VLp, displayed the strongest connections with the orbital frontal lobe. We failed to detect any thalamic regions that were most strongly connected with the anterior temporal lobe. A supplementary analysis of 3 ROIs that excluded the medial temporal ROI revealed strong connectivity of the anterior temporal lobe with the medial portion of the thalamus (results not shown). These findings, again, suggest a substantial overlap between the connectivities of the lesion sites with the medial temporal lobe and the anterior temporal lobe.

4. Discussion 4.1. Disconnected fiber bundles in anterior thalamic infarcts This study demonstrated that anterior thalamic infarcts caused widespread disconnections of the thalamocortical networks even though the lesions were circumscribed and very small; the seedonly analysis showed massive disruptions of the thalamo-frontal and thalamo-temporal fibers and of fibers to the medial parietal lobe, cerebellum and upper brainstem. These results are consistent with

previous reviews of thalamic infarcts and our previous PET study, in which the functions of the frontal and temporal cortices were predominantly affected by anterior thalamic infarcts (Carlesimo et al., 2011; Nishio et al., 2011; Schmahmann, 2003; Van der Werf et al., 2000). Although the connections between the anterior portion of the thalamus, the posterior cortices, brainstem and cerebellum have not been given much attention in human studies, these fiber connections have been demonstrated in previous connectional neuroanatomical studies in animals (Jones, 2007; Schmahmann & Pandya, 2006). We conducted seed-and-target and connectivity-based seed segmentation analyses to examine the fiber connections between the thalamus and the frontal and temporal lobes. Disruptions of the following three fiber bundles were identified: (1) the MTT; (2) the anterior, inferior and superior thalamic peduncles; and (3) the temporopulvinar bundle of Arnold. In agreement with our findings, nearly all previous case reports regarding anterior thalamic infarcts reported the involvement of the MTT. The anterior and inferior thalamic peduncles are massive fiber bundles that contain neural fibers that broadly connect the thalamus with the frontal and temporal cortices. Several previous studies have suggested that thalamic infarcts may dissect the fibers conveyed by the anterior and inferior thalamic peduncles at the location at which they merge with the IML within the

270

Y. Nishio et al. / Neuropsychologia 53 (2014) 264–273

Fig. 2. Results of the seed-only analysis of tractography. The red areas indicate fiber pathways that were disrupted in the patients with left anterior thalamic infarcts. The lesion sites are indicated in cyan.

Fig. 3. The mammillothalamic tract was delineated as a disrupted fiber tract in the seed-only analysis. The mammillary nucleus is indicated by a yellow circle.

thalamus (Graff-Radford et al., 1990; Mori et al., 1986; Nishio et al., 2011; von Cramon et al., 1985). However, this interpretation is not applicable to the numerous cases in which the anterior thalamic infarcts spared the IML. Our connectivity-based seed segmentation analysis demonstrated that the thalamo-frontal and thalamo-temporal connections, which are largely conveyed by the anterior and inferior thalamic peduncles, were disrupted at the region of the VA proper, which is affected in nearly all cases of anterior thalamic infarcts, rather

than the IML (Carlesimo et al., 2011; Van der Werf et al., 2000). Although the VA proper has been named as a ‘nucleus’, this region is rich in neural fibers and contains a relatively sparse cellular component. The fibers conveyed by the thalamic peduncles massively penetrate through the VA proper, which gives this region a lobulated appearance in myelin-stained preparations (Jones, 2007). Therefore, damage to the VA proper appears to result in diverse damage to the thalamic peduncles and produces a similar disconnection mechanism to that proposed for capsular genu infarction (Tatemichi et al., 1992). The third fiber bundle identified in this study was the temporopulvinar bundle of Arnold, which conveys fibers from the anterior temporal lobe and parahipocampal gyrus to the posterior portion of the thalamus (Schmahmann & Pandya, 2006).To the best of our knowledge, no previous studies have highlighted the involvement of this fiber tract in anterior or paramedian thalamic infarcts, probably because the regions affected by these thalamic infarcts are distant from the pulvinar. Although information concerning the fibers that are carried by the temporopulvinar bundle is currently very limited in humans and animals, Saunders and Aggleton (2007) reported that the majority of the fibers that arise from the medial temporal lobe and travel through the temporopulvinar bundle terminate in the pulvinar and the lateral dorsal nucleus, and a subset of these fibers reach the AN in macaque monkeys. Our tractography analyses revealed a pathway that was located medially in the thalamus and appeared to connect the lesion sites with the postero-medial portion of the thalamus. One possible interpretation is that the tractographically delineated fibers that are conveyed by the temporopulvinar bundle pass through the IML and reach the AN (Supplementary figure). Another possible interpretation is that the delineation of the temporopulvinar bundle is simply an artifact of fiber tracking errors. Further detailed neuroanatomical information concerning intrathalamic fiber connections, which is currently unavailable, is required to determine which of these interpretations is correct.

Y. Nishio et al. / Neuropsychologia 53 (2014) 264–273

271

Fig. 4. Results of the seed-and-target tractography analysis. A, Cartoon of the thalamic nuclei and related fiber bundles. B, Connections between the lesion sites and the left middle frontal (cyan in transverse sections) and orbital frontal (dark blue in transverse sections) target regions of interest (ROIs) are shown in purple and green, respectively, in the rendered (left) and transverse section (right) images. The area shown in cyan in the rendered image indicates the thalamus. C, Maps showing the connections between the thalamic lesion seed ROIs and the medial temporal lobe (cyan in transverse sections) and anterior temporal lobe (dark blue in transverse sections) target ROIs. The fiber pathways to the medial temporal lobe and to the anterior temporal lobes are indicated in yellow and red, respectively. The thalamus is shown in cyan in the rendered image (left). Abbreviations: AN, anterior nuclei; IML, internal medullary lamina; LD, lateral dorsal nucleus; VA, ventral anterior nuclei; VL, ventral lateral nuclei; LP, lateral posterior nucleus; VPL, ventral posterior lateral nucleus; PULV, pulvinar nuclei; and MD, mediodorsal nuclei.

Table 3 Connectivity of subregions of the thalamus with other brain areas. Thalamic regions Nomenclature of Hirai and Jones

Nomenclature of Hassler

VA proper VAmc

Lpo Lpo.mc Zo Voa Vop Doe Voi Doi Mfa Lam

VLa VLp MD IML/CeM R MTT ICg

MTL/ ATL

MFG OFL

(þ ) (þ ) () () () () () () (þ ) (þ ) (þ ) (þ ) ()

(þ) () (þ) (þ) (þ) () () (þ) () () (þ) () (þ)

( þ) ( þ) () ( þ) () ( þ) ( þ) ( þ) () () ( þ) ( þ) ( þ)

Shaded rows indicate the thalamic structures that were affected in all 6 patients. Abbreviations: VA, ventral anterior nucleus; VAmc, magnocellular ventral anterior nucleus; VLa, ventral lateral anterior nucleus; VLp, ventral lateral posterior nucleus; VM, ventral medial nucleus; MD, mediodorsal nuclei; IML, internal medullary lamina; CeM, central medial nucleus; R, reticular nucleus; MTT, mammillothalamic tract; ICg, genu of the internal capsule; MTL, medial temporal lobe; ATL, anterior temporal lobe; MFG, middle frontal gyrus; and OFL, orbital frontal lobe.

4.2. Multiple thalamo-medial temporal disconnections and memory impairment The Papez or Delay–Brion circuit, which arises from the hippocampus via the fornix, mammillary body and MTT and then projects to the AN, is thought to play a pivotal role in memory function (Gaffan & Wilson, 2008; Mayes, 2000). Previous clinico-anatomical studies have demonstrated that the MTT is involved in nearly all cases of anterior thalamic infarcts, and these findings have led us to the view that amnesia associated with this condition may be attributable to disruption of the Papez circuit (Carlesimo et al., 2011; Van der Werf et al., 2000). However, this view contrasts with the evidence from neuropsychological studies in animals. Isolated damage to a single component structure of the Papez circuit (e.g., the AN, fornix or hippocampus) or damage to the rhinal-MD-prefrontal circuit produces only mild-to-moderate memory impairment, whereas concurrent disruption of these two circuits results in dense amnesia (Gaffan & Parker, 2000; Gaffan et al., 2001; Mishkin, 1978; Murray & Mishkin, 1998; Parker & Gaffan, 1997). Based on these findings from animal studies, the dual-/multiple-system theories,

which posit that the Papez and rhinal-MD-prefrontal circuits work in concert to support normal memory function, have been proposed (Aggleton & Brown, 1999, 2006; Mishkin, 1982; Zola-Morgan & Squire, 1993). The role of rhinal-MD disconnection has been suggested in several previous studies of thalamic infarct-associated amnesia (Graff-Radford et al., 1990; Mori et al., 1986; Nishio et al., 2011; von Cramon et al., 1985). However, this mechanism was considered to be applicable to only a subset of cases because the authors of the relevant studies believed that rhinal-MD disconnection occurs at the IML. Our tractography analyses demonstrated that the inferior thalamic peduncle, which conveys most of the rhinal-MD connection fibers (Saunders et al., 2005), might be dissected at the VA proper, which is involved in nearly all cases of anterior thalamic infarcts (Carlesimo et al., 2011; Van der Werf et al., 2000). We argue that most anterior thalamic infarcts result in multiple disruptions of the thalamo-medial temporal circuits and that the observed amnesia that is associated with this condition is better understood in the framework of dual/multiple-system theories. We believe that our results fill the gap between the interpretations of human thalamic amnesias and animal neuropsychological studies. 4.3. Thalamo-cortical networks for language/semantic processing In agreement with our previous PET studies of anterior thalamic infarcts (Nishio et al., 2011), a seed-only analysis of tractography demonstrated disruptions of the fiber connections from the thalamus to the anterior temporal and lateral frontal cortices. These cortical regions are thought to be involved in lexical and semantic processing. Based on the results of our connectivity-based seed segmentation analysis, we propose three disconnection mechanisms for the language/lexico-semantic deficits associated with anterior thalamic infarcts. First, as in cases of memory impairment, the dissection of the anterior and inferior thalamic peduncles at the VA proper appear to substantially effect language functions because these fiber tracts convey the neural fibers that connect several thalamic nuclei with the temporal and frontal cortices. The second mechanism involves damage to the VAmc. In monkeys, this nucleus projects to the TE, which is a putative homolog of the human anterior temporal lobe (Middleton & Strick, 1996). Third, damage to the cellular component of the VA proper and the anterior portions of the VLa and VLp may lead to dysfunction of the lateral frontal cortices. These nuclei have abundant connections with the lateral frontal cortices (Fang, Stepniewska, & Kaas, 2006; Jones, 2007). An intracerebral EEG study by Wahl and colleagues demonstrated that event-related potentials associated with semantic

272

Y. Nishio et al. / Neuropsychologia 53 (2014) 264–273

operations are not observable from bipolar electrodes within the Vim (the posterior VLp in the terminology of Jones (2007)) but are detectable from monopolar Vim electrodes with mastoid references through a volume conduction mechanism. Based on these findings, these authors proposed possible contributions of the intralaminar nuclei, which are located adjacent to the Vim, and their projections to the prefrontal cortex (Raymer, Moberg, Crosson, Nadeau, & Rothi, 1997; Wahl et al., 2008). Alternatively, we propose that thalamically evoked potentials associated with semantic processing may originate from the VAmc, VA proper or anterior portions of the VLa and VLp, all of which are located immediately anterior to the Vim. Although, to the best of our knowledge, no human studies have addressed the connection between the thalamus and anterior temporal lobe, a recent tractography study demonstrated connections between the left frontal operculum (Broca's area) and the anterior ventral thalamus (Ford et al., 2013). Previous clinical and functional neuroimaging studies have suggested that various thalamic and cortical structures cooperate flexibly in a variety of language processing tasks (Klostermann, Krugel, & Ehlen, 2013; Llano, 2013). However, empirical evidence for the roles of individual thalamic nuclei in specific language functions, e.g., articulation, phonological processing and lexicosemantic processing, is scarce at present. These issues should be addressed in future neuroimaging and invasive neurophysiological studies in humans. 4.4. Frontal dysfunction and its contribution to memory and language deficits The thalamic nuclei form multiple neural circuits together with the basal ganglia and the frontal cortices. Three categories of circuits are associated with cognition and behavior: (i) the lateral prefrontal circuits engaged in executive cognitive function; (ii) the orbitofrontal circuits engaged in behavioral and affective control; and (iii) the anterior cingulate circuits associated with motivation and autoactivation (Cummings, 1993; Middleton & Strick, 2001). Our tractography analyses demonstrated that all of these thalamoprefrontal circuits were disrupted in anterior thalamic infarcts. In agreement with this finding, our patients exhibited cognitive and behavioral disturbances that have been linked to frontal dysfunctions, e.g., executive dysfunction and apathy (Nishio et al., 2011). Executive control processes play a critical role in the retrieval of episodic and semantic memories (Badre & Wagner, 2007). In addition, volition or motivation is prerequisite for a variety of executive, strategic cognitive processes. Therefore, prefrontal dysfunctions associated with disruptions of the thalamo-frontal circuits may have a substantial impact on memory and language functions in patients with anterior thalamic infarcts. 4.5. Limitations Several limitations of the current study should be mentioned. First, the sample size of this study was quite small. As isolated anterior thalamic infarction is a relatively rare condition, the recruitment of a large number of patients for a single study is difficult. Meta-analyses of studies that employed detailed neuroimaging analyses or neuropathological confirmations would be beneficial. Second, the precision of lesion localization using MRI is limited by image distortion due to magnetic field inhomogeneity and inaccuracies in spatial normalization and image coregistration. Third, although connectional anatomical studies have demonstrated that the ventromedial frontal regions are densely connected with the DM through the inferior thalamic peduncle (Jones, 2007; Schmahmann & Pandya, 2006), we did not evaluate this region because of susceptibility artifacts in diffusion MRI. Fourth, we did not perform tractography analyses of the brains of the patients. Negative fiber-tracking

that is based on alternate samples of healthy brains may lead to an overestimation of the fiber disconnections that actually occur in the brains of patients. Finally, because more than 7 years were required to recruit the patients of this study, we did not update the neuropsychological tests. Therefore, we could not incorporate important theories of the dissociable cognitive roles of different thalamomedial temporal networks, such as the recollection/familiarity components of memory.

Acknowledgements This work was supported by a Grant-in-Aid for Scientific Research (B) (24390278 to E.M.) and a Grant-in-Aid for Scientific Research for Young Scientists (90451591 to Y.N.) and a Grant-in-Aid for Scientific Research for Young Scientists (B) (90451591 to Y.N.).

Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.neuropsychologia. 2013.11.025. References Aggleton, J. P. (2008). EPS mid-career award 2006. Understanding anterograde amnesia: Disconnections and hidden lesions. Quarterly Journal of Experimental Psychology (Colchester), 61(10), 1441–1471. Aggleton, J. P., & Brown, M. W. (1999). Episodic memory, amnesia, and the hippocampal-anterior thalamic axis. Behavioral and Brain Sciences, 22(3), 425–444 (discussion 444-489). Aggleton, J. P., & Brown, M. W. (2006). Interleaving brain systems for episodic and recognition memory. Trends in Cognitive Sciences, 10(10), 455–463. Badre, D., & Wagner, A. D. (2007). Left ventrolateral prefrontal cortex and the cognitive control of memory. Neuropsychologia, 45(13), 2883–2901. Behrens, T. E., Berg, H. J., Jbabdi, S., Rushworth, M. F., & Woolrich, M. W. (2007). Probabilistic diffusion tractography with multiple fibre orientations: What can we gain? NeuroImage, 34(1), 144–155. Behrens, T. E., Johansen-Berg, H., Woolrich, M. W., Smith, S. M., Wheeler-Kingshott, C. A., Boulby, P. A., et al. (2003). Non-invasive mapping of connections between human thalamus and cortex using diffusion imaging. Nature Neuroscience, 6(7), 750–757. Bender, L., & Schilder, P. (1933). Encephalopathia alcoholica (polioencephalitis h ae morrhagica superior of wernicke). Archives of Neurology and Psychiatry, 29(5), 990–1053. Binder, J. R., & Desai, R. H. (2011). The neurobiology of semantic memory. Trends in Cognitive Sciences, 15(11), 527–536. Brett, M., Leff, A. P., Rorden, C., & Ashburner, J. (2001). Spatial normalization of brain images with focal lesions using cost function masking. NeuroImage, 14(2), 486–500. Carlesimo, G. A., Lombardi, M. G., & Caltagirone, C. (2011). Vascular thalamic amnesia: A reappraisal. Neuropsychologia, 49(5), 777–789. Carrera, E., & Bogousslavsky, J. (2006). The thalamus and behavior: Effects of anatomically distinct strokes. Neurology, 66(12), 1817–1823. Ciccarelli, O., Behrens, T. E., Altmann, D. R., Orrell, R. W., Howard, R. S., JohansenBerg, H., et al. (2006). Probabilistic diffusion tractography: A potential tool to assess the rate of disease progression in amyotrophic lateral sclerosis. Brain, 129 (Pt 7), 1859–1871. Cummings, J. L. (1993). Frontal-subcortical circuits and human behavior. Archives of Neurology, 50(8), 873–880. De Witte, L., Brouns, R., Kavadias, D., Engelborghs, S., De Deyn, P. P., & Marien, P. (2011). Cognitive, affective and behavioural disturbances following vascular thalamic lesions: A review. Cortex, 47(3), 273–319. Eichenbaum, H., Yonelinas, A. P., & Ranganath, C. (2007). The medial temporal lobe and recognition memory. Annual Review of Neuroscience, 30, 123–152. Fang, P. C., Stepniewska, I., & Kaas, J. H. (2006). The thalamic connections of motor, premotor, and prefrontal areas of cortex in a prosimian primate (Otolemur garnetti). Neuroscience, 143(4), 987–1020. Ford, A. A., Triplett, W., Sudhyadhom, A., Gullett, J., McGregor, K., Fitzgerald, D. B., et al. (2013). Broca's area and its striatal and thalamic connections: A diffusionMRI tractography study. Frontiers in Neuroanatomy, 7, 8, http://dx.doi.org/ 10.3389/fnana.2013.00008. Fukatsu, R., Fujii, T., Tsukiura, T., Yamadori, A., & Otsuki, T. (1999). Proper name anomia after left temporal lobectomy: A patient study. Neurology, 52(5), 1096–1099. Gaffan, D., & Parker, A. (2000). Mediodorsal thalamic function in scene memory in rhesus monkeys. Brain, 123(Pt 4), 816–827.

Y. Nishio et al. / Neuropsychologia 53 (2014) 264–273

Gaffan, D., Parker, A., & Easton, A. (2001). Dense amnesia in the monkey after transection of fornix, amygdala and anterior temporal stem. Neuropsychologia, 39(1), 51–70. Gaffan, D., & Wilson, C. R. (2008). Medial temporal and prefrontal function: Recent behavioural disconnection studies in the macaque monkey. Cortex, 44(8), 928–935. Gainotti, G. (2000). What the locus of brain lesion tells us about the nature of the cognitive defect underlying category-specific disorders: A review. Cortex, 36(4), 539–559. Graff-Radford, N. R., Tranel, D., Van Hoesen, G. W., & Brandt, J. P. (1990). Diencephalic amnesia. Brain, 113(Pt 1), 1–25. Gutman, D. A., Holtzheimer, P. E., Behrens, T. E., Johansen-Berg, H., & Mayberg, H. S. (2009). A tractography analysis of two deep brain stimulation white matter targets for depression. Biological Psychiatry, 65(4), 276–282. Harding, A., Halliday, G., Caine, D., & Kril, J. (2000). Degeneration of anterior thalamic nuclei differentiates alcoholics with amnesia. Brain, 123(Pt 1), 141–154. Jones, E. G. (2007). The thalamus (2nd ed.). Cambridge; New York: Cambridge University Press. Klostermann, F., Krugel, L. K., & Ehlen, F. (2013). Functional roles of the thalamus for language capacities. Frontiers in Systems Neuroscience, 7, 32, http://dx.doi.org/ 10.3389/fnsys.2013.00032. Kopelman, M. D. (1995). The Korsakoff Syndrome. British Journal of Psychiatry, 166, 154–173. Kopelman, M. D., Thomson, A. D., Guerrini, I., & Marshall, E. J. (2009). The Korsakoff syndrome: Clinical aspects, psychology and treatment. Alcohol and Alcoholism, 44(2), 148–154. Levin, N., Ben-Hur, T., Biran, I., & Wertman, E. (2005). Category specific dysnomia after thalamic infarction: A case-control study. Neuropsychologia, 43(9), 1385–1390. Llano, D. A. (2013). Functional imaging of the thalamus in language. Brain and Language, 126(1), 62–72. Malamud, N., & Skillicorn, S. A. (1956). Relationship between the Wernicke and the Korsakoff Syndrome—A Clinicopathologic Study of 70 Cases. Archives of Neurology and Psychiatry, 76(Dec), 585–596. Mayes, A. R. (2000). Effects on memory of Papez cicuit lesions. In: 2nd ed. L. S. Cermak (Ed.), Memory and its disorders, Vol. 2 (pp. 111–131). Amsterdam: Elsevier. Middleton, F. A., & Strick, P. L. (1996). The temporal lobe is a target of output from the basal ganglia. Proceedings of the National Academy of Sciences of the United States of America, 93(16), 8683–8687. Middleton, F. A., & Strick, P. L. (2001). A revised neuroanatomy of frontal-subcortical circuits. In: D. G. LIchter, & J. L. Cummings (Eds.), Frontal-subcortical circuits in psychiatric and neurological disorders (pp. 44–58). New York: Guilford Press. Mishkin, M. (1978). Memory in monkeys severely impaired by combined but not by separate removal of amygdala and hippocampus. Nature, 273(5660), 297–298. Mishkin, M. (1982). A memory system in the monkey. Philosophical Transactions of the Royal Society of London, Series B, 298(1089), 83–95. Molko, N., Cohen, L., Mangin, J. F., Chochon, F., Lehericy, S., Le Bihan, D., et al. (2002). Visualizing the neural bases of a disconnection syndrome with diffusion tensor imaging. Journal of Cognitive Neuroscience, 14(4), 629–636. Mori, E., Yamadori, A., & Mitani, Y. (1986). Left thalamic infarction and disturbance of verbal memory: A clinicoanatomical study with a new method of computed tomographic stereotaxic lesion localization. Annals of Neurology, 20(6), 671–676. Murray, E. A., & Mishkin, M. (1998). Object recognition and location memory in monkeys with excitotoxic lesions of the amygdala and hippocampus. Journal of Neuroscience, 18(16), 6568–6582. Nishio, Y., Hashimoto, M., Ishii, K., & Mori, E. (2011). Neuroanatomy of a neurobehavioral disturbance in the left anterior thalamic infarction. Journal of Neurology, Neurosurgery & Psychiatry, 82(11), 1195–1200. Nowinski, W. L., Thirunavuukarasuu, A., & Benarbid, A. L. (2005). The Cerefy Clinical Brain Atlas. Ojemann, G. A. (1977). Asymmetric function of the thalamus in man. Annals of the New York Academy of Sciences, 299, 380–396. Ojemann, G. A., & Ward, A. A., Jr. (1971). Speech representation in ventrolateral thalamus. Brain, 94(4), 669–680.

273

Paller, K. A., Acharya, A., Richardson, B. C., Plaisant, O., Shimamura, A. P., Reed, B. R., et al. (1997). Functional neuroimaging of cortical dysfunction in alcoholic Korsakoff's syndrome. Journal of Cognitive Neuroscience, 9(2), 277–293. Parker, A., & Gaffan, D. (1997). The effect of anterior thalamic and cingulate cortex lesions on object-in-place memory in monkeys. Neuropsychologia, 35(8), 1093–1102. Patterson, K., Nestor, P. J., & Rogers, T. T. (2007). Where do you know what you know? The representation of semantic knowledge in the human brain. Journal of Cognitive Neuroscience, 8(12), 976–987. Pergola, G., Bellebaum, C., Gehlhaar, B., Koch, B., Schwarz, M., Daum, I., et al. (2013). The involvement of the thalamus in semantic retrieval: A clinical group study. Journal of Cognitive Neuroscience, http://dx.doi.org/10.1162/jocn_a_00364. Pitel, A. L., Chetelat, G., Le Berre, A. P., Desgranges, B., Eustache, F., & Beaunieux, H. (2012). Macrostructural abnormalities in Korsakoff syndrome compared with uncomplicated alcoholism. Neurology, 78(17), 1330–1333. Rai, M., Okazaki, Y., Inoue, N., Araki, K., Fukunaga, R., & Sawada, T. (2004). Object use impairment associated with left anterior thalamic infarction. European Neurology, 52(4), 252–253. Raymer, A. M., Moberg, P., Crosson, B., Nadeau, S., & Rothi, L. J. (1997). Lexicalsemantic deficits in two patients with dominant thalamic infarction. Neuropsychologia, 35(2), 211–219. Reed, L. J., Lasserson, D., Marsden, P., Stanhope, N., Stevens, T., Bello, F., et al. (2003). FDG-PET findings in the Wernicke–Korsakoff syndrome. Cortex, 39(4-5), 1027–1045. Saksida, L. M., & Bussey, T. J. (2010). The representational-hierarchical view of amnesia: Translation from animal to human. Neuropsychologia, 48(8), 2370–2384. Saunders, R. C., & Aggleton, J. P. (2007). Origin and topography of fibers contributing to the fornix in macaque monkeys. Hippocampus, 17(5), 396–411. Saunders, R. C., Mishkin, M., & Aggleton, J. P. (2005). Projections from the entorhinal cortex, perirhinal cortex, presubiculum, and parasubiculum to the medial thalamus in macaque monkeys: Identifying different pathways using disconnection techniques. Experimental Brain Research, 167(1), 1–16. Schmahmann, J. D. (2003). Vascular syndromes of the thalamus. Stroke, 34(9), 2264–2278. Schmahmann, J. D., & Pandya, D. N. (2006). Fiber pathways of the brain. Oxford; New York: Oxford University Press. Scoville, W. B., & Milner, B. (2000). Loss of recent memory after bilateral hippocampal lesions. 1957. Journal of Neuropsychiatry & Clinical Neurosciences, 12(1), 103–113. Segal, J. B., Williams, R., Kraut, M. A., & Hart, J., Jr. (2003). Semantic memory deficit with a left thalamic infarct. Neurology, 61(2), 252–254. Semenza, C., Mondini, S., & Zettin, M. (1995). The anatomical basis of proper name processing. A critical review. Neurocase, 1(2), 183–188. Simmons, W. K., & Martin, A. (2009). The anterior temporal lobes and the functional architecture of semantic memory. Journal of the International Neuropsychological Society, 15(5), 645–649. Tatemichi, T. K., Desmond, D. W., Prohovnik, I., Cross, D. T., Gropen, T. I., Mohr, J. P., et al. (1992). Confusion and memory loss from capsular genu infarction: A thalamocortical disconnection syndrome? Neurology, 42(10), 1966–1979. Thompson-Schill, S. L. (2003). Neuroimaging studies of semantic memory: Inferring “how” from “where”. Neuropsychologia, 41(3), 280–292. Van der Werf, Y. D., Witter, M. P., Uylings, H. B., & Jolles, J. (2000). Neuropsychology of infarctions in the thalamus: A review. Neuropsychologia, 38(5), 613–627. von Cramon, D. Y., Hebel, N., & Schuri, U. (1985). A contribution to the anatomical basis of thalamic amnesia. Brain, 108(Pt 4), 993–1008. Voon, V., Kubu, C., Krack, P., Houeto, J. L., & Troster, A. I. (2006). Deep brain stimulation: Neuropsychological and neuropsychiatric issues. Movement Disorders, 21(Suppl 14), S305–327. Wahl, M., Marzinzik, F., Friederici, A. D., Hahne, A., Kupsch, A., Schneider, G. H., et al. (2008). The human thalamus processes syntactic and semantic language violations. Neuron, 59(5), 695–707. Wixted, J. T., & Squire, L. R. (2011). The medial temporal lobe and the attributes of memory. Trends in Cognitive Sciences, 15(5), 210–217. Yonelinas, A. P., Aly, M., Wang, W. C., & Koen, J. D. (2010). Recollection and familiarity: Examining controversial assumptions and new directions. Hippocampus, 20(11), 1178–1194. Zola-Morgan, S., & Squire, L. R. (1993). Neuroanatomy of memory. Annual Review of Neuroscience, 16, 547–563.