Article Text

Download PDFPDF

Research paper
Aberrant interhemispheric homotopic functional and structural connectivity in amyotrophic lateral sclerosis
  1. Jiuquan Zhang1,2,
  2. Bing Ji2,
  3. Jun Hu3,
  4. Chaoyang Zhou1,
  5. Longchuan Li2,4,
  6. Zhihao Li5,
  7. Xuequan Huang1,
  8. Xiaoping Hu6
  1. 1Department of Radiology, Southwest Hospital, Third Military Medical University, Chongqing, P.R. China
  2. 2Biomedical Imaging Technology Center, Emory University/Georgia Institute of Technology, Atlanta, Georgia, USA
  3. 3Department of Neurology, Southwest Hospital, Third Military Medical University, Chongqing, P.R. China
  4. 4Marcus Autism Center, Children's Healthcare of Atlanta, Emory University, Atlanta, Georgia, USA
  5. 5Institute of Affective and Social Neuroscience, Shenzhen University, Guangdong, P.R. China
  6. 6Department of Bioengineering, University of California, Riverside, California, USA
  1. Correspondence to Professor Xuequan Huang, Department of Radiology, Southwest Hospital, Third Military Medical University, Chongqing 400038, P.R. China; huang_xq66{at} or Professor Zhihao Li, Institute of affective and Social Neuroscience, Shenzhen University, Shenzhen, 518060, Guangdong, P.R. China; zhihao_li{at}


Objective Amyotrophic lateral sclerosis (ALS) is an idiopathic and fatal neurodegenerative disease of the human motor system. While microstructural alterations in corpus callosum (CC) have been identified as a consistent feature of ALS, studies directly examining interhemispheric neural connectivity are still lacking. To shed more light on the pathophysiology of ALS, the present study aims to examine alterations of interhemispheric structural and functional connectivity in individuals with ALS.

Methods Diffusion tensor imaging (DTI) and resting-state functional MRI (rfMRI) data were acquired from 38 individuals with ALS and 35 gender-matched and age-matched control subjects. Indices of interhemispheric functional and structural neural connection were derived with analyses of voxel mirrored homotopic connectivity (VMHC) and probabilistic fibre tracking.

Results The rfMRI has revealed extensive reductions of VMHC associated with ALS in brain regions of the precentral and postcentral gyrus, the paracentral lobule, the superior temporal gyrus, the middle cingulate gyrus, the putamen and the superior parietal lobules. With DTI, the analysis has also revealed reductions of interhemispheric structural connectivity through the CC subregions II, III and V in patients with ALS. Additionally, interhemispheric functional connectivity of the bilateral precentral gyri positively correlated with fractional anisotropy values of the CC subregion III, which structurally connects the bilateral motor cortices.

Conclusions The present data provided direct evidence confirming and extending the view of impaired interhemispheric neural communications mediated by CC, providing a new perspective for examinations and understanding the pathophysiology of ALS.

View Full Text

Statistics from


Amyotrophic lateral sclerosis (ALS) is an idiopathic and fatal neurodegenerative disease of the human motor system characterised by progressive muscular weakness and atrophy.1 Previous neuroimaging studies have identified the involvement of the motor corpus callosum (CC) to be a consistent feature of ALS.2 Particularly, tract-based spatial statistics with diffusion tensor imaging have shown a consistent reduction of fractional anisotropy (FA) in the CC extending rostrally and bilaterally to the regions of the primary motor cortex.2 In addition, it has been reported that FA values in motor-related regions of CC are reduced more than those in other CC areas in individuals with ALS.3 ,4

Given that CC is the largest white matter structure interconnecting the two cerebral hemispheres, with more than 300 million fibres,4 structural alterations in CC are expected to be associated with compromised interhemispheric functional communications. Indeed, using resting-state functional MRI (rfMRI), an overall decrease in functional connectivity between bilateral motor cortices was also noted in individuals with limb-onset ALS.5 Parallel to alterations in functional connectivity, neuroimaging studies have also revealed white matter degenerations associated with ALS.6–8 Specifically, the bilateral precentral gyri and some other abnormal interhemispheric pathways have been revealed.7

The previous findings of interhemispheric connections in ALS,5 ,7 although insightful and informative, are limited in two aspects: (1) only connections of motor cortices were examined while potential effects in other cerebral regions remain unexplored;5 and (2) structural7 and functional5 connectivity were examined separately; thus, multimodal confirmations are still lacking. To achieve a relatively more comprehensive view of the effect on interhemispheric functional and structural connections, the present study was performed with multimodal MRI in three levels. (1) With rfMRI, we examined whole-brain interhemispheric functional connectivity in ALS using a recently validated analysis approach of voxel mirrored homotopic connectivity (VMHC).9 The VMHC quantifies functional connectivity between each voxel in one hemisphere and its mirrored counterpart in the opposite hemisphere. Characterising the intrinsic functional architecture of the brain, this method has recently been successfully applied in studies of different psychiatric or neurological disorders.10 ,11 (2) With diffusion tensor imaging (DTI), we examined alterations of interhemispheric structural connectivity mediated by five CC subregions defined with an established parcellation scheme.12 (3) With the imaging measurements derived in (1) and (2), we explored possible correlations between the structural and functional imaging metrics as well as correlations between imaging metrics and clinical scores.



Thirty-eight patients (25 men) with the diagnosis of sporadic probable or definite ALS according to the revised El Escorial criteria were recruited. All patients were assessed and assigned a score for ‘ALS Functional Rating Scale-Revised’ (ALSFRS-R) within 12 hours after the MRI. Clinical variables of ‘disease duration’, measuring the temporal extent from the symptom onset to the scanning date and ‘rate of disease progression’, defined as (48—ALSFRS-R)/(disease duration), were also obtained. All patients included in the present study had not received any specific treatment. Exclusion criteria were: (1) family history of motor system diseases; (2) clinical diagnosis of frontotemporal dementia; (3) major psychiatric or neurological disorders; and (4) cognitive impairment (Montreal Cognitive Assessment score <=26).

Thirty-five healthy controls (21 men) with no history of neurological or psychiatric conditions and with normal brain MRI were recruited from the local community. All participants were right-handed based on measurements of the Edinburgh inventory.

This study was approved by the Medical Research Ethics Committee of the Southwest Hospital. Written informed consent was obtained from all participants.

MRI acquisition

MRI data were obtained using a 3T Siemens Tim Trio (Siemens, Erlangen, Germany) scanner. The DTI data were acquired using a single-shot twice-refocused spin echo sequence with repetition time (TR)=10 000 ms, echo time (TE)=92 ms, 64 diffusion directions (b=0 and 1000 s/mm2), matrix=128×124, field of view (FOV)=256×248 mm2, and 75 axial slices (thickness=2 mm, gap=0 mm). Each participant was scanned twice with the DTI for increased signal-to-noise ratio. For structural MRI, a three-dimensional (3D) magnetisation-prepared rapid gradient echo sequence was applied for sagittal images with TR=1900 ms, TE=2.52 ms, inversion time=900 ms, flip angle (FA)=9°, matrix=256×256, thickness=1.0 mm, gap=0 mm, 176 slices and voxel size=1×1×1 mm3. Finally, rfMRI data were acquired using an echoplanar imaging (EPI) sequence with 36 axial slices, slice thickness=3 mm, gap=1 mm, TR=2000 ms, TE=30 ms, FA=90°, FOV=192×192 mm2, matrix=64×64, voxel size=3×3×3 mm3, and total volumes=240. During the resting state scans, subjects were instructed to keep their eyes closed, remain still, not to think of anything in particular and not to fall asleep.

The present DTI analysis of interhemispheric structural connectivity involved probabilistic fibre tracking through different subregions of CC. For this purpose, an independent sample was needed for the unbiased parcellation of CC. This independent sample was obtained from the Human Connectome Project (HCP;, an open-access repository of healthy human brain datasets (Van Essen et al, 2012). Subjects of the HCP were scanned with a customised 3T Siemens scanner (Connectome Skyra) using a single-shot, single refocusing spin-echo, EPI sequence. Eighty subjects (32 men; ages: 22–36 years) in the HCP data set of ‘WU-Minn’ (The Washington University and University of Minnesota) were used in this study. The detailed imaging protocol can be found at

Data analysis

VMHC analysis (rfMRI)


Preprocessing of rfMRI data was performed using the toolbox of Data Processing Assistant for Resting-State fMRI (DPARSF) implemented in SPM8 ( The first 10 volumes were discarded to allow for magnetisation to reach steady state and adaptation of the participants to the scanning environment. The preprocessing steps included slice timing correction, volume registration, spatial normalisation (Montreal Neurological Institute (MNI) space), spatial smoothing (full width at half maximum (FWHM)=4 mm),9 ,10 spurious variances (signals of linear drift, head motion, ventricle, white matter and whole brain) reduction and band-pass filtering (0.01–0.08 Hz). Finally, functional images of each subjects were registered to a study-specific and symmetric MNI template for computing the VMHC.

VMHC analysis

The analysis of VMHC was also performed using the DPARSF software. For each subject, the homotopic functional connectivity was computed as the Pearson correlation coefficient between each voxel's preprocessed signal time series and that of its symmetrical counterpart in the other hemisphere. Correlation coefficients were then Fisher's z-transformed to improve normality. The resultant z-values, constituting the VMHC, were used for subsequent voxel-wise group comparison.

Using the VBM8 toolbox (, anatomical images were bias-corrected, tissue classified and registered using linear (12-parameter affine) and nonlinear transformations, within a unified model. Then, the grey matter (GM) partitions were modulated to derive local measurements of GM volumes. The GM volume was treated as a covariate of no interest in all analyses of voxel-wise group comparisons.

DTI data analysis

(The detailed steps involved in the DTI data processing were depicted in the online supplementary figure S1.)


Our DTI data were preprocessed using the FMRIB's Diffusion Toolbox (FDT) included in the FMRIB's Software Library (FSL, The preprocessing steps for DTI data included the correction of eddy current distortion, brain extraction, diffusion tensor fitting and estimation of diffusion orientation distribution functions (dODF).

DTI data of the independent HCP sample was obtained and preprocessed using HCP's diffusion preprocessing pipeline.

Masks of CC and cerebral cortex

The DTI-based fibre tracking involved 3D masks in the CC and cerebral cortex. The mask of the CC was defined in the MNI space based on the Jülich probabilistic histological atlas included in FSL (figure 1A). In the cortex, five masks were adopted in each hemisphere, including the bilateral (1) prefrontal cortex, (2) premotor cortex, (3) motor cortex, (4) somatosensory cortex and (5) parietal cortex, occipital cortex and temporal cortex (figure 1B).

Figure 1

Connectivity-based CC parcellation. (A) CCs were generated from each participant's T1 image. (B) The five cortical regions used for DTI tractography. (C) CC subregions derived from the tractography. CC, corpus callosum; DTI, diffusion tensor imaging.

CC segmentation with HCP diffusion data

With diffusion data from the independent sample of HCP, probabilistic tractography (5000 samples per voxel) was performed for the purpose of CC segmentation. This analysis accommodated crossing fibres in each voxel12 and the tracking was from all voxels in CC to the five cortical regions in both hemispheres. As a result, a probabilistic map from CC to each of the five pairs of cortical masks was generated. Then, different CC voxels in each participant were classified into five classes according to the cortical region they mostly connected to (winner-take-all).13 After this classification at an individual level, the algorithm of ‘Simultaneous Truth and Performance Level’ (STAPLE) was applied to fuse individual segmentations.14 The STAPLE approach was used to combine manual tracings of the same subject from multiple raters and to estimate the hidden ‘ground truth’ consensus from a probabilistic model.

After fusing individual segmentations, the merged result (figure 1C) was transformed from the MNI space into the individual diffusion space for subsequent DTI analysis of interhemispheric structural connectivity.

Interhemispheric fibre tractography

Here, probabilistic fibre tracking was performed twice between each pair of conjugate cortical masks with one being the seed and the other being the target and vice versa. The corresponding subregion of CC was used as the waypoint mask in these tractographies. For each subject, crossing fibres modelling was applied and the resultant value of each voxel in the tractographic images represented the number of fibre projections passed through. These projection numbers were normalised by the product of voxel number in the seed mask and the total tracks of 5000. A subject-level probability map was derived by the mean of left-to-right and right-to-left tracking results. These probability maps were further clipped at different thresholds (5%, 10%, 25% and 50%) for the visualisation (figure 2) and extraction of DTI metrics.

Figure 2

Interhemispheric homotopic structural connectivity of the five cortical regions at different probabilistic thresholds of 5%, 10%, 25% and 50%.

DTI metrics

With white matter pathways defined by the tractography, DTI measures, including fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD) and radial diffusivity (RD) were extracted from the mask of these binarised pathways.

Statistical analysis

All non-voxel-wise statistical analyses were performed with SPSS (Released 2008. SPSS Statistics for Windows, V.17.0. Chicago, Illinois, USA). χ2 and independent samples t-tests were used for discrete and continuous variables, respectively.

Group comparisons of VMHC were conducted using ‘REST’ software ( using a voxel-wise two-sample t-test controlling for age and gender. A Monte Carlo simulation ( was used to estimate the cluster size for the correction of multiple comparisons. A corrected significance level of p=0.05 was achieved with a combination of a voxel-wise threshold of p<0.01 and a cluster size threshold of volume >486 mm3.

To explore the possible correlations between the structural and functional imaging measures that showed a group difference, as well as between these imaging measures and clinical variables, Pearson correlation analyses were performed between each DTI (FA, AD, RD and MD)-fMRI (VMHC) pair2 and imaging-clinical variable pair.


Demographic and clinical data

The demographic and clinical characteristics of the present sample are summarised in table 1. All individuals with ALS had clinical signs of upper and lower motor neuron involvement. Seven individuals with ALS were classified as ‘bulbar onset’, 30 as ‘limb onset’ and 1 as ‘both bulbar and limb onsets’.15 There was no significant difference in age and gender between the ALS and control participants.

Table 1

Demographic data of the participants

Group differences in VMHC

As shown in figures 3 and 4, and table 2, ALS-associated VMHC reductions were observed in the precentral gyrus, the postcentral gyrus, the paracentral lobule, the superior temporal gyrus, the middle cingulate gyrus, the putamen and the superior parietal lobule.

Table 2

Regions showing aberrant interhemispheric functional connectivity between groups

Figure 3

Statistical maps showing a voxel-wise group comparison of VMHC. The colour bar indicates significant contrast T values of ALS-control. ALS, amyotrophic lateral sclerosis; VMHC, voxel mirrored homotopic connectivity.

Figure 4

Spatial correspondences between findings of interhemispheric structural (red-yellow) and functional (blue) connectivity. Connections through different subregions of CC are shown in different columns. Radiological convention, left is right. CC, corpus callosum.

Group differences on interhemispheric structural connectivity

Tractography results of interhemispheric structural connectivity are shown in figure 2 for the five cortical regions. At the probabilistic threshold of 25%, impaired interhemispheric structural connectivity through CC subregions II, III, and V were observed in the ALS group (figures 4 and 5). Specifically, decreased FA and increased MD/RD were noted in the fibre bundles connecting the bilateral premotor and motor cortices through CC subregions II and III, respectively; decreased FA and increased AD/RD were also noted in the fibre bundles connecting the bilateral POTC through the CC subregion V.

Figure 5

Group comparisons of diffusion metrics (FA, MD, AD and RD) for each cortical region at the probabilistic thresholds of 25% (** p<0.001, * p<0.05). AD, axial diffusivity; FA, fractional anisotropy; MD, mean diffusivity; RD, radial diffusivity.

Correlation analysis

The VMHCs of interhemispheric motor cortices positively correlated with FA values in the fibre bundles connecting bilateral motor cortices through CC subregion III(r=0.39, p<0.001) (figure 6). However, between imaging metrics and clinical scores, no correlation reached statistical significance after the Bonferroni correction for multiple comparisons.

Figure 6

Significant correlation between interhemispheric functional and structural connectivity for the bilateral motor cortex. FA, fractional anisotropy; VMHC, voxel mirrored homotopic connectivity.


The involvement of CC has been increasingly identified as one of the core features of ALS. Specifically, structural neuroimaging has shown an ALS-associated decrease of fractional anisotropy and an increase in matched regional radial diffusivity,2 particularly in the CC subregions of II, III and V.3 Additionally, neuropathological studies have shown prominent and consistent CC involvement in ALS, most abundant in the middle part.16 ,17 Neuroelectrophysiological studies such as TMS and MEG have also provided indirect evidence of impaired CC functioning by showing alterations of interhemispheric inhibition18 and cortical β-band oscillations.19 Based on these structural and functional alterations of CC, the present study complemented/extended these previous studies by showing a compromised interhemispheric neural connectivity. Namely, the consequences of CC alterations in ALS have been revealed by the present observations of extensively decreased interhemispheric functional connectivity and impaired interhemispheric structural connectivity through CC subregions II, III and V. In addition, alterations of functional and structural interhemispheric connectivity were found to be correlated for the bilateral precentral gyri. To highlight and assist in understanding the detailed brain structures involved, interhemispheric structural and functional alterations observed are visualised in a combined fashion in figure 4 with spatial correspondences of these measurements depicted. Notably again, decrease of interhemispheric functional connectivity in the patients corresponded well to the areas where they had impaired interhemispheric structural connectivity through CC subregion III.

In this study, the analysis of VMHC was applied for the first time to investigate alterations of interhemispheric functional connectivity in ALS. Individuals with ALS exhibited lower VMHC in the precentral and postcentral gyri, the paracentral lobule, the superior temple gyrus, the middle cingulate gyrus, the putamen and the superior parietal lobule. These reductions indicate an extensive disruption of interhemispheric functional connectivity in ALS.

Alteration of the sensorimotor network is the most prominent neuroimaging feature in ALS.20 ,21 In motor-related cortices, the precentral gyrus is the key region supporting complex motor behaviour. A previous study had reported sensorimotor functional connectivity changes20 and specific cortical thinning in the precentral gyrus in ALS.22 Independent component analysis has also revealed less functional connectivity of the sensorimotor network in ALS.21 The superior parietal lobe is critical for sensorimotor integration by maintaining an internal representation of the body's state. A previous study23 observed atrophy of the superior parietal gyrus in half of their ALS sample. The paracentral lobule is mainly concerned with motor and sensory innervations of the contralateral lower extremity and the regulation of physiological functions, such as defaecation and micturition. A pathological study documented that a large number of degenerating fibres can be seen to pass to or from the precentral gyrus and paracentral lobule in patients with ALS.16 A structural motor connectome analysis revealed that patients with ALS demonstrated a significantly impaired structural network overlapping of the bilateral primary motor regions (precentral gyrus and paracentral lobule, Brodmann area 4), bilateral supplementary motor regions (caudal middle frontal gyrus, BA 6), parts of the left basal ganglia (pallidum) and right posterior cingulate and precuneus.24 Collectively, the present results have added new evidence of the involvement of the sensorimotor cortex in ALS from the perspective of interhemispheric homotopic functional connectivity.

Aberrant homotopic functional connectivity in the bilateral putamen was also found in the present study. Traditionally, ALS was considered as a pyramidal system disorder with a pathogenicity mainly involving the corticospinal tracts. However, mounting evidence has indicated the involvement of extramotor areas in ALS. For example, a positron emission tomography (PET) study reported hypometabolism in the motor-sensory cortex and in the putamen.25 By using a multiple kernel learning approach, ALS-deviant functional connectivity was characterised by hyper-connected networks spanning non-cortical motor areas (notably putamen and cerebellum) as well as extramotor cortical areas. These hyper-connected networks showed compromised connectivity to the motor cortices.26 Consistent with the present results, all of these multimodal findings suggest the involvement of putamen in ALS, which challenges the traditional view that ALS is solely a pyramidal disorder.

We also observed a VHMC reduction in the superior temporal gyrus. A previous task fMRI with word retrieval reported an ALS-associated abnormal activation pattern in the middle and superior temporal gyri.27 A DTI study also found increased MD in the right superior temporal gyrus of patients with ALS.28 Additionally, a PET study revealed a reduction of regional cerebral blood flow in the bilateral superior temporal cortex, which was associated with verbal fluency dysfunction.27 Based on our findings and these previous reports, we speculate that the compromised interhemispheric functional connectivity in the superior temporal gyrus may contribute to the language impairment in ALS.29

Impaired interhemispheric structural connectivity through CC subregions II, III and V was also observed in the ALS group, and the corresponding cortices are the bilateral premotor cortices, the motor cortices and the combined parietal, occipital and temporal cortices, respectively. Previous studies2 ,3 have reported that the involvement of CC in ALS is mainly located in the motor-related subregions II and III. However, in addition to subregions II and III, the present study also noted reductions of homotopic structural connectivity through CC subregion V. In a disease that mainly affects the motor system, impairment of homotopic structural connectivity through the motor-related CC subregions are well expected, but reduced structural connectivity through CC subregion V in ALS has also been previously reported.3 Given that ALS is not simply a motor but a multisystem disorder30–33 that involves the visual cortex and is potentially mediated by parieto-occipital regions,4 the impairment of homotopic structural connectivity through CC subregion V provided additional evidence in support of an extensive pathogenicity beyond the motor circuit.

As shown in figure 4, although the present results of interhemispheric functional and structural connectivity are mostly convergent, a divergence was noted. Namely, impaired interhemispheric functional connectivity of the postcentral gyri is supposed to be mediated by CC subregion IV, whereas impaired interhemispheric structural connectivity was not observed in this subregion, but in subregion II instead. This discrepancy may be due to two reasons: (1) the analyses of rfMRI and DTI data were at different spatial resolutions, with rfMRI being voxel-wise but DTI being region of interest (ROI)-wise; and (2) instead of direct pathways through CC, interhemispheric functional connectivity could be mediated by indirect structural connections (ie, via a third region).34

Accumulating evidence indicates that neurodegenerative disease, such as Alzheimer's disease (AD), Parkinson's disease (PD) and ALS, have common cellular and molecular mechanisms.35 Given the proposed common pathogenicity of AD, PD and ALS, we suggest that in addition to ALS, the present findings may also indicate potential implications in other neurodegenerative disorders, such as AD and PD. When similar disruptions of neural connectivity have been confirmed in AD and PD, more evidence will support the hypothesis that ‘ALS, PD and AD are a spectrum of the same disorder’.

Although the group average showed a generally decreased neural connectivity in ALS, some patients did exhibit a strong connection (see individual data in figure 6). In fact, this larger inhomogeneity is a common phenomenon observed in different clinical stages of ALS;8 ,36–39 so it may reflect a more general and stage-dependent issue. Therefore, improvement of the present diagnostic scheme may involve more biological features and instead of a general ‘disordered’ group, future studies may consider subcategorising the clinical stages of ALS.38

Interhemispheric functional and structural connectivity correlated significantly with the bilateral precentral gyri. This finding correlates well with previous reports that the motor CC is more affected than other CC subregions in ALS.2 ,3 Furthermore, the correlation may suggest that alterations of the interhemispheric structural connectivity through CC subregion III may subserve the aberrant interhemispheric functional connectivity of the bilateral precentral gyri.

In the present study, a significant correlation was not observed between imaging measurements and clinical variables. Verstraete et al40 noted this issue, reviewed potential contributing factors and concluded that the presence of such a clinical imaging correlational gap may result partly from the limitations of clinical assessments, partly from the limitations of imaging methodology and partly from the inherent biology of this complex disease.

The limitations of the present study involve three aspects. First, the human brain is not completely symmetrical. However, the analysis of VMHC assumes and enforces a symmetrical standard template. Thus, the present study is insensitive to potential asymmetric group differences. Second, due to concerns about subject compliance, our sample size was relatively small as only individuals with mild and moderate symptoms were included. Third, the lack of a comprehensive evaluation of non-motor symptoms weakened the interpretation and exploration of brain–behaviour relationships.


View Abstract


  • JQZ and BJ contributed equally to this study.

  • Contributors JQZ, ZHL, XQH and XPH conceived and designed the experiments. JQZ, JH and CYZ performed the experiments. JQZ and BJ analysed the data. JQZ, BJ, LCL and ZHL contributed to reagents/materials/analysis tools. JQZ and BJ contributed by writing the manuscript. JQZ, JH and CYZ collected the data.

  • Funding This work was supported by funds from the National Natural Science Foundation of China (grant number 81200882, 31671169), the Natural Science Foundation of Chongqing (grant number CSTC2016jcyjA2163) and the Natural Science Foundation of SZU (grant number 201564, 000099).

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval The Medical Research Ethics Committee of the Southwest Hospital.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles