ORIGINAL RESEARCH published: 21 April 2020 doi: 10.3389/fneur.2020.00240 The Free and Cued Selective Reminding Test in Parkinson’s Disease Mild Cognitive Impairment: Discriminative Accuracy and Neural Correlates Edited by: Mayela Rodríguez-Violante, Andrea Horta-Barba 1,2,3,4 , Javier Pagonabarraga 1,2,3,4*, Saül Martínez-Horta 1,2,3,4 , Instituto Nacional de Neurología y Juan Marín-Lahoz 1,3,4 , Frederic Sampedro 1,2,3 , Ramón Fernández-Bobadilla 1,2 , Neurocirugía Manuel Velasco M. Ángeles Botí 1,2 , Helena Bejr-Kasem 1,2,3 , Ignacio Aracil-Bolaños 1,2 , Suárez, Mexico Jesus Pérez-Pérez 1,2,3 , Berta Pascual-Sedano 1,3,4 , Antonia Campolongo 2,3 , Reviewed by: Cristina Izquierdo 2 , Beatriz Gómez-Ansón 2,4,5 and Jaime Kulisevsky 1,2,3,4 Alexandru Hanganu, Centre de Recherche de l’Institut 1 Movement Disorders Unit, Neurology Department, Sant Pau Hospital, Barcelona, Spain, 2 Biomedical Research Institute Universitaire de Gériatrie de Montréal (IIB-Sant Pau), Barcelona, Spain, 3 Centro de Investigación en Red-Enfermedades Neurodegenerativas (CIBERNED), (CRIUGM), Canada Barcelona, Spain, 4 Autonomous University of Barcelona, Barcelona, Spain, 5 Neuroradiology Unit, Sant Pau Hospital, Jordi A. Matias-Guiu, Barcelona, Spain Hospital Clínico San Carlos, Spain *Correspondence: Introduction: Memory alterations are common in Parkinson’s disease (PD) patients but Javier Pagonabarraga jpagonabarraga@santpau.cat the mechanisms involved in these deficits remain poorly understood. The study aims to explore the profile of episodic memory deficits in non-demented early PD patients. Specialty section: This article was submitted to Methods: We obtained neurological, cognitive and behavioral data from 114 PD patients Movement Disorders, and 41 healthy controls (HC). PD participants were grouped as normal cognition (PD-NC) a section of the journal Frontiers in Neurology and mild cognitive impairment (PD-MCI) according to the Level II criteria of the Movement Disorders Society Task Force (MDS-TF). We evaluate the performance amongst groups Received: 19 November 2019 Accepted: 12 March 2020 on an episodic memory task using the Free and Cued Selective Reminding Test (FCSRT). Published: 21 April 2020 Additionally, gray matter volume (GMV) voxel based morphometry, and mean diffusivity Citation: (MD) analyses were conducted in a subset of patients to explore the structural brain Horta-Barba A, Pagonabarraga J, Martínez-Horta S, Marín-Lahoz J, correlates of FCSRT performance. Sampedro F, Fernández-Bobadilla R, Results: Performance on all subscores of the FCSRT was significantly worse in PD-MCI Botí MÁ, Bejr-Kasem H, Aracil-Bolaños I, Pérez-Pérez J, than in PD-NC and HC. Delayed total recall (DTR) subscore was the best at differentiating Pascual-Sedano B, Campolongo A, PD-NC from PD-MCI. Using crosstabulation, DTR allowed identification of PD-MCI Izquierdo C, Gómez-Ansón B and Kulisevsky J (2020) The Free and patients with an accuracy of 80%. Delayed free and cued recall was associated with Cued Selective Reminding Test in decreased GMV and increased MD in multiple fronto-temporal and parietal areas. Parkinson’s Disease Mild Cognitive Impairment: Discriminative Accuracy Conclusion: Encoding and retrieval deficits are a main characteristic of PD-MCI and and Neural Correlates. are associated with structural damage in temporal, parietal and prefrontal areas. Front. Neurol. 11:240. doi: 10.3389/fneur.2020.00240 Keywords: PD-MCI, Free and Cued Selective Reminding Rest, Parkinson’s disease, memory, episodic memory Frontiers in Neurology | www.frontiersin.org 1 April 2020 | Volume 11 | Article 240 Horta-Barba et al. The FCSRT in Parkinson’s Disease INTRODUCTION METHODS At the time Parkinson’s disease (PD) diagnosis, up to 30% of Participants patients meet diagnostic criteria for mild cognitive impairment We prospectively recruited 114 PD patients who fulfilled the UK (PD-MCI) (1–3). The rate of progression of PD-MCI is Brain Bank Diagnostic Criteria for PD and regularly attending heterogeneous, with up to 36% of patients fulfilling diagnostic the Movement Disorders Unit at our center and a group criteria for PD dementia (PDD) after 4 years after diagnosis, of 41 age-matched and education-matched healthy controls. with cumulative PDD prevalence of 80% in 20 years long- The study procedures included a neurological examination term survivors (4–6). Thus, despite dementia is not an and the administration of a comprehensive neuropsychological inevitable consequence of PD, it affects a significant proportion assessment battery which was done to all participants, including of patients for which treatments to ameliorate this entity patients and healthy controls. Presence of PDD according are lacking. to consensus guidelines (12); having undergone deep brain Identifying early cognitive indicators suggestive of stimulation surgery; brain abnormalities evidenced in imaging progression to dementia is a major need to stratify patients studies performed in the previous year; major depression; in different groups of risk and also to design interventions treatment with anticholinergic drugs; and any known causes of before PDD onset. In this sense, the addition of posterior- cognitive impairment other than PD defined exclusion criteria. cortical type deficits -to the prototypical frontal-executive Written informed consent was obtained from all participants alterations seen in most PD patients- seem to characterize and all procedures were performed in accordance with the the transition from PD-MCI to PDD in this population. standards of the local Ethic Review Board of the Sant Pau hospital Accordingly, the development of language, memory and in Barcelona, and with the 1964 Helsinki declaration and its visuospatial/visuoperceptive alterations are indicative of a later amendments. more aggressive progression of cognitive deterioration in PD (7, 8). Procedures Episodic memory alterations are also found in PD and Data was collected during two separate visits. Data at screening affects up to 45% of de novo PD patients. However, its included: age, educational level, current medications with role in the delineation of progression from PD-MCI to PDD dopaminergic drugs converted to levodopa equivalent daily dose and the mechanisms participating in these deficits has been (LEDD), formal application of the MDS criteria to exclude scarcely studied (9). Attention and retrieval deficits -rather PDD, the MDS-Unified Parkinson’s Disease Rating Scale part III than storage and retention alterations- has been pointed to (UPDRS-III) motor subscale, Hoehn, and Yahr (H&Y), and the sub-serve episodic memory difficulties in PD (10). This is Parkinson’s Disease-Cognitive Rating Scale (PD-CRS), which is supported by the benefit commonly observed in retrieval when a screening instrument that addresses global cognition. On the semantic or recognition cues are presented to PD patients. second visit, a comprehensive neuropsychological examination Accordingly, decreased performance in memory tasks in PD that fulfilled the standards proposed by the MDS Task Force for have been attributed to frontal-executive deficits rather than the diagnosis of PD-MCI was completed (2). to hippocampal or medial temporal lobe alterations. However, difficulties in retrieving information even during recognition Neuropsychological Assessment and Group and cued-facilitated recall have also been described in PD Classification (11). This suggests that in some patients amnestic difficulties PD participants were grouped as PD-NC and PD-MCI according may be associated to hippocampal alterations rather than been to the Level II criteria of the Movement Disorders Society restricted to frontal-executive alterations. The role that this kind Task Force (MDS-TF) for the diagnosis of PD-MCI (1, 2). of deficits might play in PD-MCI and in the conversion to Thus, five cognitive domains (attention, language, memory, PDD is mostly unknown. However, exploring differences in visuospatial skills, and executive functions) were examined using episodic memory performance in non-demented PD patients a comprehensive battery composed of two tests per domain. with and without PD-MCI may help to delineate early cognitive We applied cut-offs of 1.5 SD below normative values and changes with significant prognostic implications in terms of PD-MCI was confirmed when any two (or more) impaired cognitive progression. neuropsychological test were present (2, 13). The following In the present study, we aimed to explore the profile of standardized and recommended neuropsychological measures episodic memory deficits in non-demented early PD patients were used: Parkinson’s Cognitive dementia rating scale (PD- with normal cognition (PD-NC) and PD-MCI. To explore CRS), forward and backward Corsi’s block-tapping task, forward the extent of structural brain differences accompanying these digit span task, phonetic and semantic verbal fluency, the Rey- deficits, we also conducted voxel based morphometry (VBM) Osterrieth complex figure test, the Boston Naming test, the and mean diffusivity (MD) analyses in a subset of participants. Judgment of Line Orientation, and the number location subtest Gray matter volume (GMV) analyses through VBM is a of the Visual Object and Space Perception Battery. macrostructural neuroimaging technique that has been widely Episodic memory was assessed using the Free and Cued used to characterize brain atrophy. In recent years, increases Selective Reminding Test (FCSRT). The FCSRT is a widely used in MD both in white-matter and gray-matter tissues have been episodic memory test which assesses immediate and delayed suggested to infer microstructural brain damage. free-recall and cued-facilitated immediate and delayed recall. In Frontiers in Neurology | www.frontiersin.org 2 April 2020 | Volume 11 | Article 240 Horta-Barba et al. The FCSRT in Parkinson’s Disease studies assessing memory performance in MCI in the general >0.6, large effect size). Receiver operator characteristic (ROC) population, the FCSRT has shown to reflect hippocampal- curves were generated to explore the discriminative capacity mediated consolidation memory defects better than free-list of each FCSRT subscore. We used cross-tabulation to calculate learning tests. Moreover, the FCSRT performance predicts diagnostic accuracy. Associations between demographic, clinical progression to dementia in close relationship with progressive and cognitive variables were studied using Pearson’s correlations. atrophy of the medial temporal lobe and other neocortical Significance was set at p < 0.05. All the statistical procedures were temporal and parietal regions (10, 14–19). The FCSRT was performed using the SPSS v16.0 statistical software package. administered using standard procedures as described by Grober and Buschke (20). Participants were shown a card with four Neuroimaging Analysis words, and were asked to determine which one of the four A voxel-based morphometry (VBM) analysis of gray matter corresponds to a particular category (e.g., cue; clothing, and volume (GMV) was performed using the Statistical Parametrical the word “vest”). The participant should learn the four items Mapping (SPM12) software (http://www.fil.ion.ucl.ac.uk/spm/). on the four cards (total 16 words). Three recall trials were T1-MRI images were segmented to obtain GMV probability conducted, each one preceded by 20 s of counting backwards maps, which were then normalized to the Montreal Neurological used as interference. For each trial, participants were asked Institute (MNI) stereotactic space using DARTEL. The resulting to freely recall as many items as possible and category cues images were smoothed using a Gaussian kernel of 8 mm full were provided for items not retrieved by total free recall. The width at half maximum (FWHM). same procedure of recalling (freely and cued) was done after a DTI images were preprocessed using FSL 5.0 software 30 min interval. Subjects were required to freely remember the (https://fsl.fmrib.ox.ac.uk/fsl/fslwiki). First, non-brain tissue was words and category cues were provided for items not retrieved removed using the Brain Extractor Tool (BET). Second, motion freely. The measures evaluated here were: total free recall-TFR and eddy current correction was performed using the FMRIB’s (cumulative sum of free recall from the three trials; range 0–48), Diffusion Toolbox (FDT). Diffusion tensors were then computed total recall-TR (cumulative sum of free recall + cued recall from and mean diffusivity (MD) maps were obtained for each patient. the three trails, range 0–48), delayed free recall-DFR (free delayed These maps were then normalized to MNI space and smoothed recall, range 0–16), and delayed total recall-DTR (free delayed using an isotropic filter of 6 mm FWHM. recall + cued delayed recall, range 0–16). The normalized GMV and MD images were entered into Both patients and healthy controls followed the a voxel-wise multiple regression analysis to explore the brain same assessment. correlates of the DFR and DTR scores in both modalities. Age, sex, education and total intracranial volume were used as Neuroimaging Acquisition covariates of no interest. Voxelwise imaging results showing p < A subsample of 56 patients underwent 3-Tescla Magnetic 0.005 (uncorrected) and a minimum cluster extent size of k = Resonance Imaging (MRI) (Philips Achieva). T1-weighted MRI 100 voxels was considered significant (21–23). Clusters surviving acquisition was performed using a dedicated axial T13D- family-wise error (FWE) correction for multiple comparisons are MPRAGE MRI (TR/TE, 500/50 ms; flip angle, 8, field of view reported in the corresponding cluster description table (Table 4). [FOV], 23 cm; with in-plane resolution of 256 × 256 and 1- The MRIcron software tool (https://www.nitrc.org/projects/ mm slice thickness). Diffusion Tensor Imaging (DTI) scans were mricron) was used to represent the statistical voxelwise maps. also obtained (FOV 220 mm, voxel size 2 mm, TR 8,000 ms, TE 80 ms, flip angle 90◦ , 32 directions, b-factor 1000). The RESULTS neuropsychological and MRI scans were performed within a maximum of 3 months between procedures. One hundred and fourteen PD patients (68.0 ± 8.3 years) and 41 healthy controls (HC; 66.3 ± 7.4 years) were included in the Statistical Analysis study. PD patients were in the early to mid-stages of the disease Data are expressed as means ± standard deviation (SD) for the (disease duration 5.3 ± 3 years; H&Y stage 2.0 ± 0.2; UPDRS-III continuous variables and as mean range for the ordinal variables. 25.2 ± 8.1). As seen in Table 1, PD patients and HC were matched Differences between groups were analyzed with independent for age, gender and education. two-tailed t-tests and analyses of variance (ANOVA) for According to MDS-TF level II criteria for PD-MCI, using a continuous variables, the Mann-Whitney test for ordinal data, detection threshold of −1.5 SD, 22 of the 114 patients (19.0%) and the χ2 test for categorical variables. Comparison of clinical, were classified as having PD-MCI and all HC were classified in the demographic, and neuropsychological data between groups were range of normal cognition, following this same criteria. Looking done using One-Way ANOVA between the three groups, with at the different cognitive groups, those in the PD-MCI group additional Tukey post-hoc tests for more direct comparisons were significantly older than PD-NC [t (113) = 3.174; p < 0.01] between each pair of groups. and HC [t (63) = 3.78; p < 0.01], have lower educational level than Binary logistic regression analysis was performed to test PD-NC [t (113) = −4.18; p < 0.01] and HC [t (63) = −3.66; p < the independent classification capacity of the different FCSRT 0.01], and had higher UPDRS-III score than PD-NC [t (63) = 3.10; subscores.To calculate the effect size of the differences observed p < 0.05]. between cognitive groups we used Cohen’s d coefficient (d As depicted in Table 2, between-group comparisons showed values: 0–0.3, small effect size; 0.3–0.6, moderate effect size; significant differences in all the cognitive measures with the Frontiers in Neurology | www.frontiersin.org 3 April 2020 | Volume 11 | Article 240 Horta-Barba et al. The FCSRT in Parkinson’s Disease TABLE 1 | Clinic and sociodemographic characteristics of the sample. Controls PD total sample PD-NC PD-MCI P* P† Age (years) 66.3 ± 7.4 68 ± 8.3 67.5 ± 8.1 72.9 ± 4.9 0.251 0.012 Gender (m/f) 21/20 72/42 64/28 8/14 χ 2 = 0.125 χ 2 = 0.004 Education (years) 13.2 ± 4.8 12.2 ± 4.7 13 ± 4.4 8.7 ± 4.4 0.250 <0.001 Disease duration (years) – 5.3 ± 3.4 5 ± 3.1 6.1 ± 4.2 – 0.278 MDS-UPDRS IIIa 1.2 ± 1.2 25.2 ± 8.1 24 ± 7.4 29.8 ± 9 <0.001 <0.01 H&Yb – 2 ± 0.2 2 ± 0.2 2 ± 0.2 – 0.891 Total LEDDc – 565 ± 312 571 ± 320 540 ± 282 – 0.680 a Movement Disorders Society—Unified Parkinson’s Disease Rating Scale part III. b Hoehn and Yahr stage. c Total levodopa equivalent daily dose. *P-values were determined with t-test for independent samples between healthy controls and PD. † P-values were determined with t-test for independent samples between PD-NC, and PD-MCI. TABLE 2 | Level I and Level II assessment scores. Controlsa PD-NCb PD-MCIc ANOVA Turkey’s a−b PD-CRS Total score 101.2 ± 12.9 95 ± 15.4 73 ± 7.1 <0.001 0.056; b−c <0.001 a−b Frontal-subcortical 72.3 ± 12.1 66.4 ± 14.6 44.8 ± 7 <0.001 0.057; b−c <0.001 a−b Posterior-cortical 28.9 ± 1.7 28.5 ± 1.7 27.1 ± 2 <0.001 0.594; b−c <0.005 Attention a−b Corsi Forward 5.2 ± 0.9 5.2 ± 0.9 4.1 ± 1 <0.001 0.987; b−c <0.001 a−b Digit span forward 5.5 ± 1 5.5 ± 1.1 4.6 ± 0.7 <0.01 0.947; b−c <0.005 Executive functions a−b Phonetic fluency 15.4 ± 4.6 15.5 ± 4.9 8.7 ± 3.4 <0.001 0.993; b−c <0.001 a−b Corsi Backward 4.8 ± 0.9 4.7 ± 1 3.4 ± 0.9 <0.001 0.878; b−c <0.001 Memory a−b PD-CRS delayed memory total recall 6.6 ± 1.8 6.2 ± 2.4 4.5 ±1.7 <0.001 0.660; b−c <0.001 1 a−b ROCFT −30 min 16.1 ± 5.8 13.3 ± 6.8 4.8 ± 5.3 <0.001 0.076; b−c <0.001 Language BNT-602 54.6 ± 6.7 54.7 ± 4.6 47.7 ± 5.2 <0.001 a−b 0.685; b−c <0.001 a−b Semantic fluency 20.6 ± 4.8 19 ± 5.5 12.1 ± 3.2 <0.001 0.229; b−c 0.001 Visuospatial JLOT3 23.3 ± 4.6 22.4 ± 5 15.1 ± 6.9 <0.001 a−b 0.679; b−c <0.001 VOSP—number location4 19.7 ± 1 19.6 ± 0.9 19.3 ± 1.2 0.351 a−b 0.907; b−c 0.424 1 Rey-Osterrieth complex figure test −30 min delayed recall. 2 Boston Naming Test −60 items. 3 Judgement of line orientation test. 4 Visual object and shape perception test. a−b Controls vs. PD-NC. b−c PD-NC vs. PD-MCI. exception of the number location subtest of the VOSP. Post- TR (d =1.23), DFR (d = 1.03), and DTR (d = 1.60) when hoc comparisons showed no significant differences in cognitive comparing PD-MCI with PD-NC. Large effects were also found performance between PD-NC and controls. Conversely, PD- when comparing PD-MCI with HC in all FCSRT subscores: TFR MCI performed significantly worse than PD-NC in all cognitive (d = 1.31), TR (d = 0.94), DFR (d = 1.05), and DTR (d = measures with the exception of the number location subtest of 1.14). When comparing PD-NC with HC with Cohen’s d, we the VOSP. found small effects in all FCSRT subscores: TFR (d = 0.29); Looking at the FCSRT test performance, no differences were TR (d = 0.11); DFR (d = 0.06); and DTR (d = 0.10). We found between PD-NC and HC in any of the obtained FCSRT used stepwise logistic regression analysis (forward; conditional) subscores. Conversely, performance of the patients in the PD- to determine FCSRT subscores that independently differentiated MCI group was significantly worse than performance of the HC PD-NC from PD-MCI. The variables found to be significantly and the PD-NC in all the subscores (p < 0.001). As reflected different between cognitive groups in the one-way ANOVA were by Cohen’s d, large effect sizes were found for TFR (d = 1.09), included in the analysis to assess their contribution to group Frontiers in Neurology | www.frontiersin.org 4 April 2020 | Volume 11 | Article 240 Horta-Barba et al. The FCSRT in Parkinson’s Disease discrimination. From all the variables included in the model, the sensitivity, 86%; specificity, 81% (AUC = 0.870; 95% confidence DTR (r = 549; p < 0.001) was the best differentiating PD-NC interval 0.804–0.936, p < 0.001). Using crosstabulation, we found from PD-MCI independently on age, education or UPDRS-III DTR scores identified PD-MCI patients with 80% accuracy score (see Table 3). (see Figure 1). Discriminative Capacity of the FCSRT Neuroimaging Data Receiver operating characteristics (ROC) curves to discriminate Voxel-wise multiple regression analysis between GMV and DFR between PD-MCI and PD-NC indicated that a cut-off of ≤ 20/21 and DTR respectively showed significant positive associations points on the TFR score yielded sensitivity, 73%; specificity, between FCSRT performance and GMV in multiple fronto- 77% (AUC = 0.770; 95% confidence interval, 0.663–0.878, p < temporal and parietal areas. Specifically, poorer DFR scores 0.005). A cut-off score ≤ 38/39 points on the TR score showed were associated with decreased GMV in the left mid temporal sensitivity, 80%; specificity, 69% (AUC = 0.803; 95% confidence gyrus (BA 21), the paracentral gyrus and the superior temporal interval, 0.704–0.902, p < 0.001). A cut-off score ≤6/7 on the gyrus (BA 41) (FWE corrected p < 0.05). Less restrictive criteria DFR score showed sensitivity, 80%; specificity, 70% (AUC = (uncorrected p < 0.005; k = 100) showed positive associations 0.760; 95% confidence interval 0.647–0.874, p < 0.005). The DTR with the right mid temporal gyrus, the right superior frontal score showed the best discriminative properties to differentiate gyrus and the left inferior parietal gyrus (BA 40). Performance PD-MCI from PD-NC using a cut-off ≤12/13 points, achieving in the DTR was significantly associated with GMV in the TABLE 3 | Comparative FCSRT performance between PD-MCI, PD-NC and HC. Controlsa PD-NCb PD-MCIc ANOVA Turkey’s Cohen’s d* FCSRT a−b Total free recall 26.7 ± 7.5 24.5 ± 7 16.7 ± 7 <0.001 0.300; b−c <0.001 1.09 a−b Total recall 41.6 ± 6.7 42.8 ± 5.1 35.1 ± 9.1 <0.001 0.872; b−c <0.001 1.23 a−b Delayed free recall 9.6 ± 3.1 9.4 ± 3 6.4 ± 3 <0.001 0.952; b−c <0.001 1.03 a−b Delayed total recall 14.2 ± 2.6 14.4 ± 1.8 11 ± 3 <0.001 0.878; b−c <0.001 1.60 a−b Controlsvs. PD-NC. b−c PD-NC vs. PD-MCI. *Cohen’s d for PD-NC vs. PD-MCI comparisons. FIGURE 1 | Receiver operating characteristic (ROC) curves illustrating the discriminative properties of the FCSRT. Frontiers in Neurology | www.frontiersin.org 5 April 2020 | Volume 11 | Article 240 Horta-Barba et al. The FCSRT in Parkinson’s Disease TABLE 4 | Results of the GMV voxel-based morphometry and MD-DTI analysis. Anatomical region Cluster size T value MNI coordinates (x, y, z) VOXEL-BASED MORPHOMETRY ANALYSIS OF GMV FCSRT delayed free recall Left mid temporal (BA 21) / postcentral / superior temporal (BA 41)* 2,799 4.63 −60, −26, −5 −57, −21, 26 −57, −18, 8 Right mid temporal 551 3.25 60, −39, 9 Left inferior parietal (BA 40) 498 3.76 −47, −45, 45 Right superior frontal 371 3.55 21, 6, 59 FCSRT delayed total recall Right SMA / superior frontal (BA 6)* 1,832 4.44 17, 11, 66 21, 8, 60 Left mid temporal (BA 21) 606 3.90 −60, −26, −5 MD-DTI ANALYSIS FCSRT delayed free recall Left superior temporal (BA 22) 1,212 3.72 −46, −7, 1 Left inferior frontal (BA 47) 916 3.44 −40, 14, −5 Left hippocampus 884 4.41 −36, −21, −15 Left mid temporal 509 3.69 −56, −25, −5 Left middle temporal 246 4.23 −54, −59, 1 FCSRT delayed total recall Left inferior frontal* 4,158 4.04 −45, 12, 22 Left postcentral 2,307 3.70 −58, −11, 27 Left superior temporal 1,210 3.31 −62, −16, 1 Left inferior temporal 574 3.79 −40, −18, −19 Left superior temporal pole 553 4.12 −43, 13, −23 *Cluster level FWE corrected (p < 0.05). right supplementary motor area (SMA), in the right superior discriminative properties of different subscores of the FCSRT frontal gyrus (BA 6) (FWE corrected p < 0.05) and in the and, in a subset of patients; we addressed the structural left mid temporal gyrus (BA 21) (see Table 4, Figure 2 and neuroimaging correlates of FCSRT performance by means of Supplementary Table 1). GMV-VBM and MD. Voxel-wise MD multiple regression analysis with DFR and Our results show that, among other neuropsychological DTR showed a significant association between performance measures, PD-MCI patients perform significantly worse than and MD in multiple fronto-temporal and parietal clusters PD-NC and healthy controls in the FCSRT. No significant predominantly located in the left hemisphere. Specifically, differences were seen between PD-NC and healthy controls in associations were found between worse DFR and increased MD this respect. in the left superior temporal gyrus (BA 22), the left inferior The FCSRT performance in the PD-MCI group was worse frontal gyrus (BA 47), the left hippocampus, the left mid temporal than that in PD-NC patients and controls in all the free and cued and the left middle temporal gyrus. Performance in the DTR immediate and delayed recall conditions, suggesting difficulties was significantly associated with MD in the left inferior frontal at level of encoding, consolidation and retrieval. Interestingly, (FWE corrected p < 0.05), the left postcentral, the left superior multiple measures of memory performance and specifically a temporal gyrus, the left superior temporal pole, and the left single measure of episodic memory (DTR) correctly classified inferior temporal gyrus (see Table 4 and Figure 3). as PD-MCI up to 80% of the cases. These results emphasize Voxel-wise multiple regression analysis between GMV and the notion that beyond the prototypical frontal-executive deficits MD with both IFR and ITR scores did not show any significant characterizing cognitive impairment in PD, amnestic difficulties association surviving FWE correction. are also inherent features of the cognitive changes observed in PD. Accordingly, this indicates that early to mid in the course of DISCUSSION PD-MCI, not just frontal-executive, but also amnestic difficulties are present. In the present study we assessed performance in the FCSRT Neuroimaging data showed an association between in non-demented PD and compared the performance in this widespread cortical (temporal, parietal and prefrontal) areas task between PD-MCI and PD-NC. We also explored the and FCSRT performance. Less GMV in mid temporal (BA 21), Frontiers in Neurology | www.frontiersin.org 6 April 2020 | Volume 11 | Article 240 Horta-Barba et al. The FCSRT in Parkinson’s Disease FIGURE 2 | Results of the VBM analysis. The slices show regions with significant GMV decreases in association with poorer FCSRT performance. Section (A) (top) depicts GMV associated with DFR performance; section (B) (bottom) depicts GMV associated with DTR performance. For depiction purposes results are showed with a p < 0.005 (uncorrected) and k = 100. FIGURE 3 | Results of the voxelwise MD analysis. The slices show regions with significant MD increases in association with poorer FCSRT performance. For depiction purposes results are showed with a p < 0.005 (uncorrected) and k = 100. the superior temporal, the supramarginal, the inferior parietal which extends along the parahippocampal gyrus, running from and the superior frontal gyrus was clearly associated with the anteromedial temporal lobe to the inferior parietal and delayed free recall. Conversely, a more selective involvement occipital lobes (25). The parahippocampal cingulum bundle of superior frontal regions was found for delayed total recall. is closely linked to learning and episodic memory (26–28). All these areas have complex connections that can be grouped Furthermore, microstructural white matter changes in this into (a) temporal and parietal areas more specialized in storage region have been consistently associated with episodic and processes, and (b) prefrontal areas more specialized in retrieval recognition memory deficits in amnestic MCI and early AD processes (5, 24). Altogether, our results suggest that episodic patients (29–31). memory deficits in PD-MCI are sub-served by dysfunction of In the only previous study that has analyzed the performance parieto-temporal and prefrontal-related encoding, consolidation of FCSRT in PD patients with amnestic mild cognitive and retrieval processes. Similarly, MD-DTI analyses delineated impairment, PD patients performed worse than healthy controls the involvement of a set of fronto-temporal regions including on delayed free recall, but no differences were found regarding the temporal pole and the hippocampus. All these regions are delayed cued recall scores (10). This discrepancy could be connected through the parahippocampal cingulum bundle, explained by both the smaller sample size and the lack of Frontiers in Neurology | www.frontiersin.org 7 April 2020 | Volume 11 | Article 240 Horta-Barba et al. The FCSRT in Parkinson’s Disease comparison between PD patients with and without MCI, as the tool to detect PD-MCI patients with widespread study focused on comparing amnestic mild cognitive impairment cortical alterations. with healthy controls and patients with amnestic mild cognitive impairment without PD. Although memory deficits in PD-MCI DATA AVAILABILITY STATEMENT patients are widely considered to be caused by retrieval problems, studies using comprehensive neuropsychological batteries have The datasets generated for this study are available on request to shown that memory impairment in pre-dementia stages is also the corresponding author. the consequence of encoding and storage failure (32). The ability of a memory test, such as the FCSRT, to assess both encoding and ETHICS STATEMENT retrieval deficits would explain its appropriateness for screening PD-MCI accurately. The studies involving human participants were reviewed and Recent studies in newly diagnosed and non-demented approved by Comité etico de Investigación Clinica—Sant PD patients have also underlined the relevance and early Pau. The patients/participants provided their written informed development of cortical gray matter changes and DTI-MD consent to participate in this study. alterations in hippocampal and parahippocampal structures as predictors of worsening cognition (33, 34). AUTHOR CONTRIBUTIONS This study has several limitations. First, there were fewer PD- MCI patients than PD-NC patients. However, the prevalence of AH-B, JP, SM-H, and JK conception, design, data collection, data PD-MCI is representative of the one observed in PD patients analysis, data interpretation, writing, and editing. JM-L, FS, RF-B, in the early to mid-stages of the disease. Second, we did not MB, HB-K, IA-B, JP-P, BP-S, AC, CI, and BG-A data collection include in the study the measure “trial 1 free recall of the FCSRT” and editing. which would have given us more information about encoding. Third, the fact that we did not include patients with dementia FUNDING limits our ability to see how this population performs on FCSRT. Fourth, not all the participants underwent neuroimaging, and This work was partially supported by grants from La although the number of PD patients with available neuroimaging Marató TV3 (20142910, 2014/U/477); FIS (PI14/02058, was comparable to other VBM and DTI studies, imaging data was PI15/00962); CIBERNED. lacking for the healthy control group used in this study. And fifth, only a small subset of the clusters described in the neuroimaging ACKNOWLEDGMENTS analyses survived FWE correction. To our knowledge, this is the first study to look for FCSRT The authors thank all the patients for their participation in cut-off scores in the screening of PD-MCI by using currently this study. accepted MDS-TF criteria, providing evidence that this test is highly accurate for this purpose. Furthermore, we observed SUPPLEMENTARY MATERIAL that FCSRT impairment correlates with structural changes in crucial areas of the semantic network and memory storage. The The Supplementary Material for this article can be found combination of these clinical and imaging findings supports online at: https://www.frontiersin.org/articles/10.3389/fneur. the use of this test as an appropriate neuropsychological 2020.00240/full#supplementary-material REFERENCES 5. Patterson K, Nestor PJ, Rogers TT. Where do you know what you know? The representation of semantic knowledge in the human brain. Nat Rev Neurosci. 1. Litvan I, Aarsland D, Adler CH, Goldman JG, Kulisevsky J, Mollenhauer (2007) 8:976–87. doi: 10.1038/nrn2277 B, et al. MDS task force on mild cognitive impairment in Parkinson’s 6. Hely MA, Reid WGJ, Adena MA, Halliday GM, Morris JGL. The disease: critical review of PD-MCI. Move Disord. (2011) 26:1814–24. Sydney Multicenter Study of Parkinson’s disease: the inevitability of doi: 10.1002/mds.23823 dementia at 20 years. Move Disord. (2008) 23:837–44. doi: 10.1002/mds. 2. Litvan I, Goldman JG, Tröster AI, Schmand BA, Weintraub D, Petersen 21956 RC, et al. Diagnostic criteria for mild cognitive impairment in Parkinson’s 7. Kehagia AA, Barker RA, Robbins TW. Neuropsychological and disease: movement disorder society task force guidelines. Move Disord. (2012) clinical heterogeneity of cognitive impairment and dementia in 27:349–56. doi: 10.1002/mds.24893 patients with Parkinson’s disease. Lancet Neurol. (2010) 9:1200–13. 3. Williams-Gray CH, Foltynie T, Brayne CEG, Robbins TW, Barker doi: 10.1016/S1474-4422(10)70212-X RA. Evolution of cognitive dysfunction in an incident Parkinson’s 8. Pagonabarraga J, Kulisevsky J. Cognitive impairment and disease cohort. Brain. (2007) 130:1787–98. doi: 10.1093/brain/ dementia in Parkinson’s disease. Neurobiol Dis. (2012) 46:590–6. awm111 doi: 10.1016/j.nbd.2012.03.029 4. Foltynie T, Brayne CEG, Robbins TW, Barker RA. The cognitive ability of 9. Muslimović D, Post B, Speelman JD, Schmand B. Cognitive profile of patients an incident cohort of Parkinson’s patients in the UK. The CamPaIGN study. with newly diagnosed Parkinson disease. Neurology. (2005) 65:1239–45. Brain. (2004) 127:550–60. doi: 10.1093/brain/awh067 doi: 10.1212/01.wnl.0000180516.69442.95 Frontiers in Neurology | www.frontiersin.org 8 April 2020 | Volume 11 | Article 240 Horta-Barba et al. The FCSRT in Parkinson’s Disease 10. Costa A, Zabberoni S, Peppe A, Perri R, Fadda L, Iannarelli F, 25. Wu Y, Sun D, Wang Y, Wang Y, Ou S, Hayes DJ. Segmentation of the et al. Free and cued recall memory in Parkinson’s disease associated cingulum bundle in the human brain : a new perspective based on DSI with amnestic mild cognitive impairment. PLoS ONE. (2014) 9:e86233. tractography and fiber dissection study. Front Neuroanat. (2016) 10:1–16. doi: 10.1371/journal.pone.0086233 doi: 10.3389/fnana.2016.00084 11. Das T, Hwang JJ, Poston KL. Episodic recognition memory and the 26. Bubb EJ, Metzler-Baddeley C, Aggleton JP. The cingulum bundle: anatomy, hippocampus in Parkinson’s disease: a review. Cortex. 113:191–209. function, and dysfunction. Neurosci Biobehav Rev. (2018) 92:104–27. doi: 10.1016/j.cortex.2018.11.021 doi: 10.1016/j.neubiorev.2018.05.008 12. Emre M, Aarsland D, Brown R, Burn DJ, Duyckaerts C, Mizuno Y, et al. 27. Fernández-Matarrubia M, Matías-Guiu JA, Cabrera-Martín MN, Moreno- Clinical diagnostic criteria for dementia associated with Parkinson’s disease. Ramos T, Valles-Salgado M, Carreras JL, et al. Episodic memory dysfunction Move Disord. (2007) 22:1689–707. doi: 10.1002/mds.21507 in behavioral variant frontotemporal dementia: a clinical and FDG- 13. Goldman JG, Holden S, Ouyang B, Bernard B, Goetz CG, Stebbins PET study. J Alzheimer’s Dis. (2017) 57:1251–64. doi: 10.3233/JAD- GT. Diagnosing PD-MCI by MDS task force criteria: how many 160874 and which neuropsychological tests? Move Disord. (2015) 30:402–6. 28. Jones DK, Christiansen KF, Chapman RJ, Aggleton JP. Neuropsychologia doi: 10.1002/mds.26084 Distinct subdivisions of the cingulum bundle revealed by diffusion 14. Dubois B, Feldman HH, Jacova C, DeKosky ST, Barberger-Gateau P, MRI fibre tracking : implications for neuropsychological investigations. Cummings J, et al. Research criteria for the diagnosis of Alzheimer’s disease: Neuropsychologia. (2013) 51:67–78. doi: 10.1016/j.neuropsychologia.2012. revising the NINCDS-ADRDA criteria. Lancet Neurol. (2007) 6:734–46. 11.018 doi: 10.1016/S1474-4422(07)70178-3 29. Metzler-Baddeley C, Hunt S, Jones DK, Leemans A, Aggleton 15. Papp KV, Rentz DM, Mormino EC, Schultz AP, Amariglio RE, Johnson KA. JP, O’Sullivan MJ. Temporal association tracts and the Cued memory decline in biomarker-defined preclinical Alzheimer disease. breakdown of episodic memory in mild cognitive impairment. Neurology. (2017) 88:1431–8. doi: 10.1212/WNL.0000000000003812 Neurology. (2012) 79:2233–40. doi: 10.1212/WNL.0b013e31827 16. Quenon L, Dricot L, Woodard JL, Hanseeuw B, Gilis N, Lhommel R. 689e8 Prediction of free and cued selective reminding test performance using 30. Metzler-Baddeley C, Jones DK, Steventon J, Westacott L, volumetric and amyloid-based biomarkers of Alzheimer’s disease. J Int Aggleton JP, O’Sullivan MJ. Cingulum microstructure predicts Neuropsychol Soc. (2016) 22:991–1004. doi: 10.1017/S1355617716000813 cognitive control in older age and mild cognitive impairment. 17. Sarazin M, Berr C, De Rotrou J, Fabrigoule C, Pasquier F, Legrain J Neurosci. (2012) 32:17612–9. doi: 10.1523/JNEUROSCI.3299-1 S, et al. Amnestic syndrome of the medial temporal type identifies 2.2012 prodromal AD: a longitudinal study. Neurology. (2007) 69:1859–67. 31. Sarazin M, Chauviré V, Gerardin E, Colliot O, Kinkingnéhun S, De doi: 10.1212/01.wnl.0000279336.36610.f7 Souza LC, et al. The amnestic syndrome of hippocampal type in 18. Teichmann M, Epelbaum S, Samri D, Levy Nogueira M, Michon A, Alzheimer’s disease: an MRI study. J Alzheimer’s Dis. (2010) 22:285–94. Hampel H, et al. Free and Cued Selective Reminding Test—accuracy for doi: 10.3233/JAD-2010-091150 the differential diagnosis of Alzheimer’s and neurodegenerative diseases: 32. Weintraub D, Moberg PJ, Culbertson WC, Duda JE, Stern MB. Evidence for a large-scale biomarker-characterized monocenter cohort study (ClinAD). impaired encoding and retrieval memory profiles in Parkinson disease. Cognit Alzheimer’s Dementia. (2017) 13:913–23. doi: 10.1016/j.jalz.2016.12.014 Behav Neurol. (2004) 17:195–200. 19. Wagner M, Wolf S, Reischies FM, Daerr M, Wolfsgruber S, Jessen F, et al. 33. Carlesimo GA, Piras F, Assogna F, Pontieri FE, Caltagirone C, Spalletta Biomarker validation of a cued recall memory deficit in prodromal Alzheimer G. Hippocampal abnormalities and memory deficits in Parkinson disease. Neurology. (2012) 78:379–86. doi: 10.1212/WNL.0b013e318245f447 disease: a multimodal imaging study. Neurology. (2012) 78:1939–45. 20. Grober E, Buschke H, Crystal H, Bang S, Dresner R. Screening for dementia doi: 10.1212/WNL.0b013e318259e1c5 by memory testing. Neurology. (1988) 38:900–3. doi: 10.1212/WNL.38.6.900 34. Duffau H. New insights into the anatomo-functional connectivity of the 21. Lieberman MD, Cunningham WA. Type I and Type II error concerns in semantic system: a study using cortico-subcortical electrostimulations. Brain. fMRI research: re-balancing the scale. Soc Cognitive Affective Neurosci. (2009) (2005) 128:797–810. doi: 10.1093/brain/awh423 4:423–8. doi: 10.1093/scan/nsp052 22. Martinez-Horta S, Sampedro F, Pagonabarraga J, Fernandez-Bobadilla Conflict of Interest: The authors declare that the research was conducted in the R, Marin-Lahoz J, Riba J, et al. Non-demented Parkinson’s disease absence of any commercial or financial relationships that could be construed as a patients with apathy show decreased grey matter volume in key executive potential conflict of interest. and reward-related nodes. Brain Imaging Behav. (2017) 11:1334–42. doi: 10.1007/s11682-016-9607-5 Copyright © 2020 Horta-Barba, Pagonabarraga, Martínez-Horta, Marín-Lahoz, 23. Vaquero L, Cámara E, Sampedro F, Pérez de los Cobos J, Batlle F, Fabregas Sampedro, Fernández-Bobadilla, Botí, Bejr-Kasem, Aracil-Bolaños, Pérez-Pérez, JM, et al. Cocaine addiction is associated with abnormal prefrontal function, Pascual-Sedano, Campolongo, Izquierdo, Gómez-Ansón and Kulisevsky. This is an increased striatal connectivity and sensitivity to monetary incentives, and open-access article distributed under the terms of the Creative Commons Attribution decreased connectivity outside the human reward circuit. Addict Biol. (2017) License (CC BY). The use, distribution or reproduction in other forums is permitted, 22:844–56. doi: 10.1111/adb.12356 provided the original author(s) and the copyright owner(s) are credited and that the 24. Binder JR, Desai RH, Graves WW, Conant LL. Where is the semantic system? original publication in this journal is cited, in accordance with accepted academic A critical review and meta-analysis of 120 functional neuroimaging studies. practice. No use, distribution or reproduction is permitted which does not comply Cerebral Cortex. (2009) 19:2767–96. doi: 10.1093/cercor/bhp055 with these terms. Frontiers in Neurology | www.frontiersin.org 9 April 2020 | Volume 11 | Article 240