Brain Imaging-derived Phenotypes and Stroke: A Bidirectional Mendelian Randomization Study Unveils Causal Links between Thalamic Nuclei Volume and Stroke Risk in the European Population PDF Free Download
1 / 16/16
100%
1
Brain Imaging-derived Phenotypes and Stroke: A Bidirectional Mendelian
1
Randomization Study Unveils Causal Links between Thalamic Nuclei Volume
2
and Stroke Risk in the European Population
3
4
Xuanyan Yang1
†
, Shiyao Cheng1,2
†
, Shaoli Lin1, Peiyue Su1, Xinyao Tang1, Jiani Wu1,
5
Xingchen Lin1, Si Cheng2,3*, Siyang Liu1*
6
7
8
1 School of Public Health (Shenzhen), Sun Yat-sen University, Shenzhen 518107,
9
Guangdong, China;
10
2 Department of Neurology, Beijing Tiantan Hospital, Capital Medical University,
11
Beijing 100070, China;
12
3 China National Clinical Research Center for Neurological Diseases, Beijing 100070,
13
China;
14
15
16
17
18
Xuanyan Yang and Shiyao Cheng contributed equally to this study
19
20
21
Corresponding authors:
22
Si Cheng(sicheng@ncrcnd.org.cn)
23
Siyang Liu (liusy99@mail.sysu.edu.cn)
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324441doi: medRxiv preprint
NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.
2
45
Abstract
46
Background and Aims:
47
Recent advances in brain-imaging techniques have enabled the identification of brain
48
imaging-derived phenotypes (IDPs), representing physiological brain structure.
49
Observational studies have suggested a correlation between these IDPs and stroke,
50
confounded and based on limited samples. To investigate the causal relationship
51
between IDPs and stroke and its subtypes for an in-depth mechanistic comprehension
52
of their interplay, we conducted a bidirectional two-sample Mendelian Randomization
53
(MR) study leveraging the largest-scale genome-wide association studies (GWAS) of
54
IDPs and stroke subtypes.
55
Methods:
56
We utilized GWAS summary statistics from the BIG40 dataset, which included nearly
57
3,935 IDPs among 33,224 individuals, and GIGASTROKE, which included three
58
etiological ischemia subtypes, as well as cerebral ischemia, intracerebral hemorrhage
59
stroke and overall stroke among 73,652 stroke cases and 1,234,808 controls.
60
Results:
61
In the forward MR analysis, we identified eight significant IDPs influencing the risk
62
of stroke and its subtypes after Bonferroni correction. Notably, the volume of the
63
lateral posterior thalamus in the right hemisphere exhibited a significant negative
64
association with all ischemic stroke (OR=0.79; 95% CI: 0.74 to 0.84; p=1.21e-13), all
65
stroke (OR=0.84; 95% CI: 0.79 to 0.89; p=2.45e-9), and large vessel stroke (OR=0.54;
66
95% CI: 0.43 to 0.69; p=3.01e-7). Conversely, no significant causal association was
67
observed in the reverse MR analysis.
68
Conclusion:
69
This study enhances our understanding of causality between IDPs and stroke by
70
pinpointing specific causal associations. These findings provide valuable insights into
71
the etiology of stroke, offering potential strategies for predicting and intervening in
72
stroke risk at the level of brain imaging.
73
74
Keywords: brain imaging-derived phenotypes, stroke, genetics, causal inference,
75
Mendelian randomization.
76
77
Abbreviations
78
IDP, Imaging-Derived Phenotype; MR, Mendelian Randomization; GWAS,
79
Genome-Wide Association Study; AS, All Stroke; AIS, All Ischemic Stroke; LAS,
80
Large Vessel Stroke; CES, CardioEmbolic Stroke; SVS, Small Vessel Stroke; FA,
81
fractional anisotropy; ICVF, Intra-Cellular Volume Fraction; OD, Orientation
82
Dispersion index; TSMR, two-sample Mendelian randomization; ISOVF, Isotropic or
83
free water Volume Fraction; IV, Instrumental Variables; IVW, Inverse Variance
84
Weighted; LP, Lateral Posterior nucleus of the thalamus; LD, Lateral Dorsal nucleus
85
of the thalamus; AV, Anterior Ventral nucleus of the thalamus; CL, Central Lateral
86
nucleus of the thalamus;
87
88
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324441doi: medRxiv preprint
3
89
1. Introduction
90
Stroke is a prevalent clinical manifestation of cerebrovascular disease and a
91
substantial global public health concern. It encompasses both ischemic and
92
hemorrhagic stroke, characterized by the sudden onset of brain dysfunction(Campbell
93
and Khatri, 2020). The Global Burden of Disease study reported stroke as the second
94
leading cause of death and the third leading cause of death and disability combined in
95
2019, with ischemic stroke accounting for 62.4% of all cases(Feigin et al., 2021).
96
Although reperfusion therapy has shown efficacy in reducing disability, its practice is
97
constrained by the treatment time window(Campbell et al., 2019). In the absence of
98
effective early prevention strategies, stroke continues to exert a substantial toll on
99
lives and property(Feigin et al., 2021). Consequently, it is crucial to investigate the
100
risk factors contributing to stroke development and devise new preventive strategies.
101
102
Brain image-derived phenotypes (IDPs) are digital characteristics of brain tissue
103
obtained through the analysis of brain imaging data, offering precise, reliable, and
104
quantitative information for neuroimaging research(Gong et al., 2021). Advanced
105
imaging techniques enable the prediction of diseases before symptoms manifest,
106
particularly with risk-stratified cohorts(Miller et al., 2016). Numerous observational
107
studies have identified several associations between IDPs and stroke. For instance,
108
periventricular white-matter hyperintensity volume is positively associated with
109
cardioembolic ischemic stroke(Kaffashian et al., 2016). Significant changes in cortical
110
thickness, surface area, and gray matter volume have been observed during stroke
111
recovery(Liu et al., 2023). Patients with cerebral autosomal dominant arteriopathy
112
with subcortical infarcts and leukoencephalopathy exhibit reduced fractional
113
anisotropy (FA), altered mode of anisotropy, and increased mean diffusivity(Zhang et
114
al., 2021). Stroke has also been linked to changes in intra-cellular volume fraction
115
(ICVF), orientation dispersion index (OD), and quantitative T2 values(Siemonsen et
116
al., 2009; Wang et al., 2019). Different IDPs may contribute to stroke subtypes, and
117
conversely, stroke and its subtypes can induce alterations in IDPs. However, existing
118
observational studies often suffer from small sample sizes, and some lack specific
119
detection methods, leaving many aspects of the relationship between stroke and
120
specific brain anatomy or connectivity structures unexplored.
121
122
Mendelian randomization (MR) is a statistical approach that utilizes genetic variation
123
as an instrumental variable to assess and quantify causality(Burgess et al., 2016).
124
Genome-wide association studies (GWAS) identify genetic variants associated with
125
disease risk or traits, enabling the exploration of causality through MR. Previous
126
studies have partially unraveled the causal relationship between brain image-derived
127
phenotypes (IDPs) and stroke, with a focus on brain connectivity(Jia et al., 2023; Yu
128
et al., 2023). However, these studies did not investigate the causal link between IDPs
129
related to brain anatomy structures and stroke. Furthermore, these studies have
130
primarily relied on smaller sample sizes derived from the Megastroke study (40,585
131
cases; 446,696 controls) while the recent Gigastroke has doubled the sample size with
132
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324441doi: medRxiv preprint
4
73,652 cases and 1,234,808, controls and unraveled more association signals(Mishra
133
et al., 2022).
To obtain the most statistically robust understanding of the causalities
134
between IDPs and stroke risk, we employed a two-sample Mendelian randomization
135
(TSMR) method strategy, complemented by comprehensive sensitive analyses, on the
136
largest scale GWAS summary data available to date.
137
138
2.Method
139
2.1.Data sources
140
We employed the summary statistics of brain IDPs from Smith’s GWAS
141
meta-analysis study, which encompassed 3,935 IDPs and 17,103,079 genome-wide
142
SNPs on autosomes 1-22 and the X chromosome. The study comprised a discovery
143
sample of 22,138 individuals and a replication sample of 11,086 individuals of
144
European ancestry(Smith et al., 2021). Detailed data can be accessed on the Oxford
145
Brain Imaging Genetics (BIG40) web server
146
(https://open.win.ox.ac.uk/ukbiobank/big40/). Based on observational studies
147
mentioned before, we selected 10 categories comprising a total of 2,010 IDPs. These
148
categories included 647 regional and tissue volume, 371 cortical area, 303 cortical
149
thickness, 14 regional T2*, 75 white matter (WM) tract FA, 75 WM tract diffusion
150
tensor mode (MO), 300 WM tract diffusivity, 75 WM tract ICVF, 75 WM tract OD,
151
75 WM tract isotropic or free water volume fraction (ISOVF). Supplementary Table
152
2 provides additional information about the data sources for the IDPs.
153
154
In our study, we utilized the largest GWAS meta-analysis conducted by the
155
GIGASTROKE consortium to investigate stroke. This analysis included over one
156
million participants, with approximately 70% of European ancestry(Mishra et al.,
157
2022). Comprehensive data can be accessed through the GWAS catalog
158
(https://www.ebi.ac.uk/gwas/). Stroke was categorized into five groups: All Stroke
159
(AS), All Ischemic Stroke (AIS), CardioEmbolic Stroke (CES), Small Vessel Stroke
160
(SVS), and Large Artery Stroke (LAS). The GWAS data in the European population
161
consisted of 1,234,808 controls and varying numbers of cases for each stroke
162
subtype(Mishra et al., 2022). Supplementary Table 3 provides additional
163
information about the data sources for stroke and its subtypes.
164
165
2.2.Study design
166
To investigate the causality between IDPs and stroke, we followed a four-step process
167
based on Burgess's study for Bidirectional Two-Sample Mendelian
168
Randomization(Burgess et al., 2016). Firstly, we collected GWAS summary data from
169
public studies or websites mentioned earlier and organized the data into categories.
170
Next, we selected SNPs as instrumental variables (IVs) for each IDP and stroke type
171
based on three IV hypotheses. Thirdly, we conducted a bidirectional two-sample MR
172
using five methods separately for the causal directions of IDPs to stroke and stroke to
173
IDPs. Lastly, we performed a sensitivity analysis to validate the IVs and ensure the
174
reliability of the MR results. The workflow and a brief overview are presented in Fig.
175
1.
176
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324441doi: medRxiv preprint
5
177
Fig. 1. Workflow of the causal inference between IDPs and stroke
178
179
2.3.Instrumental variables selection
180
The three IV assumptions, as outlined by MR guidelines(Burgess et al., 2016), are as
181
follows: 1) The variants exhibit a strong association with the exposure, 2) the variants
182
are not associated with the outcome through any confounding factors, 3) the variants
183
do not directly impact the outcome, except via the exposure-outcome pathway. To
184
select independent SNPs as IVs, we employed genome-wide conditional & joint
185
association analysis (GCTA-COJO)(Yang et al., 2011). This approach evaluates the
186
variance explained by all SNPs on a chromosome or genome for a specific disease or
187
trait to ensure the independence of multiple variants at a given locus. We applied a
188
p-value threshold of 5e-8 and a linkage disequilibrium r-square less than 0.2 (gcta64
189
–bfile reference_panel_file –cojo-file gwas_file –cojo—slct –cojo-p 5e-8
190
--cojo-collinear 0.2) to select satisfactory SNPs, which were then included as IVs in
191
our subsequent study. Supplementary Table 2 provides further information on the
192
IVs for IDPs, while Supplementary Table 3 contains information on the IVs for
193
stroke. The source code and data are openly accessible in online Github repository
194
(https://github.com/liusylab/IDP_Stroke_2SMR).
195
196
2.4.Statistical analysis
197
The MR analysis process details are outlined in the STROBE checklist. At first, we
198
used the harmonise_data() function from the TwoSampleMR package to harmonize
199
the exposure and outcome data. Then, we conducted bidirectional Two Sample
200
Mendelian Randomization (TSMR) using five methods, including inverse variance
201
weighted (IVW), MR Egger, weighted mode, simple mode, and weighted median.
202
IVW was chosen as the primary method to explore the causal relationships between
203
IDPs and stroke due to its robustness(Hartwig et al., 2017), while the other four
204
methods were used to further validate the results. Sensitivity analysis was performed
205
using MR Egger to assess horizontal pleiotropy(Bowden et al., 2015), and
206
leave-one-out analysis and Cochran's Q test were also conducted. Leave-one-out
207
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324441doi: medRxiv preprint
6
analysis evaluated the reliability of the MR model, visually represented by a forest
208
plot. Cochran’s Q test was employed to identify directional heterogeneity among
209
SNPs. Additionally, MR Steiger analysis was performed to estimate the sensitivity
210
ratio and assess the correctness of the causal direction. All statistical analyses were
211
conducted using the TwoSampleMR packages in R version 4.2.2. To account for
212
multiple testing, we applied Bonferroni correction across the entire MR analysis,
213
setting the significance threshold at p<2.46e-6 (0.05/ (2010*5*2)), where 2010*5
214
represents the number of all IDP-stroke pairs and 2 denotes forward and reverse MR
215
tests.
216
217
3.Results
218
In the forward MR analysis, a total of 6,363 TSMR tests were conducted to assess the
219
causal relationship between 1,289 IDPs and 5 stroke types. Using the robust IVW
220
method, 733 results reached nominal significance (p<0.05), involving 517 IDPs (Fig.
221
2, Supplementary Table 4). Following the Bonferroni correction for multiple testing,
222
we identified eight significant causal relationships between four IDPs and three stroke
223
types. These associations were all within the same brain anatomy categories and
224
specifically related to the volume of thalamus nuclei. All the results from the five MR
225
methods, under the multiple testing correction, are provided in Fig. 3 and visualized
226
with the location and structure of the thalamus in Fig. 4.
227
228
Fig. 2. Overview of the forward MR based on IVW. (A) the result for IDPs to the
229
stroke (regional and tissue volume, cortical area, cortical thickness, WM tract ICVF),
230
(B) the result for stroke to IDPs (regional T2*, WM tract FA, WM tract diffusivity,
231
WM tract OD, WM tract ISOVF). The p-value threshold was set as 0.001 (0.05/
232
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324441doi: medRxiv preprint
7
(10*5), 10 donates a number of IDP categories, and 5 donates a number of stroke
233
types). Only associations that existed in the relationship between IDPs and stroke are
234
listed in Fig. 2.
235
236
237
Fig. 3. Causalities in the forward MR. Only results that remain significant after
238
Bonferroni correction are listed (p<2.46e-6). The arrow implies the maximum interval
239
extended on the axis.
240
241
242
Fig. 4. Schematic diagram of thalamic nuclei. Abbreviations: AN, anterior nucleus;
243
LD, lateral dorsal; LP, lateral posterior; MD, medial dorsal; VA, ventral anterior; VL,
244
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324441doi: medRxiv preprint
8
ventral lateral; VPL, ventral posterior lateral; P, pulvinar; AV, anterior ventral;AM,
245
anterior medial; AD, anterior dorsal; CL, central lateral; CM, central medial. The
246
thalamus nuclei showed no association with stroke and painted gray. The shades of LP,
247
LD, AV, and CL are linked with minimum OR.
248
249
The lateral posterior nucleus of the thalamus (LP) receives extensive afferent fibers
250
from the cerebral cortex. An increase of 1 s.d. in the volume of LP in the right
251
hemisphere was associated with a 21% lower risk of AIS (OR=0.79; 95% CI: 0.74 to
252
0.84; p=1.21e
−
13), a 16% lower risk of AS (OR=0.84; 95% CI: 0.79 to 0.89; p=2.45e
253
−
9), and a 46% lower risk of LAS (OR=0.54; 95% CI: 0.43 to 0.69; p=3.01e
−
7). The
254
lateral dorsal nucleus of the thalamus (LD) showed that an increase of 1 s.d. in its
255
volume in the right hemisphere was associated with a 21% lower risk of AIS
256
(OR=0.79; 95% CI: 0.74 to 0.84; p=5.39e
−
12) and a 15% lower risk of AS (OR=0.85;
257
95% CI: 0.80 to 0.91; p=2.37e
−
6). The anterior ventral nucleus of the thalamus (AV)
258
demonstrated that an increase of 1 s.d. in its volume in the left hemisphere was
259
associated with a 65% lower risk of LAS (OR=0.35; 95% CI: 0.26 to 0.48; p=9.02e
−
260
11) and a 30% lower risk of AIS (OR=0.70; 95% CI: 0.63 to 0.78; p=1.63e-10).
261
Similarly, the central lateral nucleus of the thalamus (CL) displayed that an increase
262
of 1 s.d. in its volume in the right hemisphere was associated with a 25% lower risk of
263
AIS (OR=0.75; 95% CI: 0.68 to 0.83; p=3.35e
−
8). Four significant IDPs are causally
264
linked to AIS, with corresponding scatter plots presented in Fig. 5. Additional scatter
265
plots for other pairs can be found in Supplementary Figure 2-5.
266
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324441doi: medRxiv preprint
9
267
Fig. 5. Scatter plots of individual SNP effects and estimates from different MR
268
methods. (A) Volume of LP in the right Thalamic Nuclei to Ischemic Stroke, (B)
269
Volume of LD in the right Thalamic Nuclei to Ischemic Stroke, (C) Volume of AV in
270
the left Thalamic Nuclei to Ischemic Stroke, (D) Volume of CL in the left Thalamic
271
Nuclei to Ischemic Stroke.
272
273
Horizontal pleiotropy was assessed using the intercept term of MR Egger regression
274
and funnel plots, which showed no evidence of its presence (Fig. 3, Supplementary
275
Figure 6-13). Directional heterogeneity was evaluated through Cochran’s Q test,
276
which revealed no evidence of such heterogeneity (Fig. 3). The leave-one-out analysis
277
demonstrated minimal change in the results after removing any individual SNP,
278
indicating the high reliability of the MR models (Supplementary Figure 14-21). In
279
summary, the sensitivity analyses confirmed the reliability of our forward MR results.
280
281
In the reverse MR analysis, we performed 8,241 tests to examine the causal
282
relationship between stroke and IDPs (Supplementary Table 5). Using IVW, we
283
identified 554 nominal significant results (p<0.05), involving 464 IDPs
284
(Supplementary Figure 1). However, after applying the Bonferroni correction, none
285
of the results remained significant.
286
287
4.Discussion
288
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324441doi: medRxiv preprint
10
Our study conducted bidirectional TSMR analyses to comprehensively investigate the
289
causal relationship between brain IDPs and stroke or its subtypes in the European
290
population. To ensure statistical validity, we utilized the largest available cohorts with
291
genetic information for IDPs and stroke. We identified several IDPs that showed
292
nominal significance with stroke and its subtypes in both causal directions (p<0.05).
293
After correcting for multiple tests, eight associations remained significant in the
294
forward MR analysis, while no results remained significant in the reverse MR
295
analysis.
296
297
In the forward MR analysis, we identified four IDPs causally associated with three
298
stroke types, all related to the volume of thalamus nuclei. LP, a thalamus nucleus
299
involved in visual processing(Casanova et al., 1991) and rotation response(Motles et
300
al., 1985), showed negative associations with AIS, AS, and LAS. In LP, stroke
301
manifestations commonly include hemihypesthesia, hemiataxia, and executive
302
dysfunction, typically resulting from artery embolism and microangiopathy(Carrera et
303
al., 2004). However, there is currently no observational evidence linking the volume
304
of LP and stroke. LD, a thalamus nucleus connected to the posterior parietal cortex
305
and cingulate gyrus(Shibata and Naito, 2005), which play roles in spatial learning and
306
memory processing(Bezdudnaya and Keller, 2008), demonstrated negative
307
associations with AIS and AS. Unfortunately, no relevant studies have explored the
308
relationship between the volume of LD and stroke. AV, a thalamus nucleus with
309
connections to the frontal cortex(Yeterian and Pandya, 1988) and cingulate
310
gyrus(Baleydier and Mauguiere, 1980), linked to behavioral activation(c-Fos
311
expression in the limbic thalamus following thermoregulatory and wake-sleep
312
changes in the rat - PubMed, n.d.) and learning functions(Shibata and Yoshiko, 2015),
313
exhibited a negative correlation between the volume of the left AV and LAS risk.
314
Although the volume of AV has been reported to decrease in psychiatric disorders like
315
schizophrenia(Forno et al., 2023), no similar study has investigated the association
316
between AV and stroke. CL, a thalamus nucleus forming part of the rostral
317
intralaminar nuclei, receiving inputs from limbic, sensorimotor-related, and cerebellar
318
input, and distributing over the frontal cortex and dorsal striatum(Vertes et al., 2022),
319
has potential functions in arousal(Xu et al., 2020) and motor control(Sakayori et al.,
320
2019). Our results indicated that a decrease in the volume of the right CL was
321
associated with a higher risk of AIS. However, no studies have measured the volume
322
of CL or explored its relationship with stroke. Previous research suggests that white
323
matter loss after thalamic infarction may derive the cortical reshaping(Conrad et al.,
324
2022). Stroke can be caused by cerebral vessels with different anatomical structures
325
through various mechanisms(Kim et al., 2019). Based on these findings, we
326
hypothesize that a genetically predicted smaller thalamic nucleus volume may result
327
in more tortuous nerve fibers and blood vessels, increasing the risk of stroke.
328
329
While the evidence regarding specific thalamus nuclei and stroke types is limited, the
330
thalamus has been consistently associated with stroke. Children with perinatal AIS
331
exhibit a smaller volumes of both thalami compared to control groups(Ilves et al.,
332
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324441doi: medRxiv preprint
11
2022). Similarly, non-thalamic stroke patients also display smaller thalamic volume
333
on both sides(Geng et al., 2023). Most observational studies have reported smaller
334
volumes in the ipsilateral or bilateral thalamus, which may partially align with our
335
findings but requires further investigation for confirmation. Assuming that the
336
subcortical grey matter remains similar before and after a stroke within six
337
weeks(Predicted Brain Age After Stroke - PubMed, n.d.), the smaller thalamic volume
338
before the stroke likely has an adverse impact by limiting resources(Liew et al.,
339
2021).
340
341
Though clinical and observational studies have indicated that stroke can lead to
342
various changes in IDPs, we did not establish a causal relationship between stroke and
343
IDPs in our reverse MR analyses. Since the imaging data was obtained from a healthy
344
population, it is expected not to observe any reverse causality. On the other hand, this
345
may suggest that the intricate mechanisms of neural connections account for the
346
absence of causality, or the effects are too subtle to detect in our study.
347
348
Our findings differ significantly from previous bidirectional TSMR findings by Yu et
349
al., published in BMC Medicine(Yu et al., 2023), and Jia et al., in Cerebral Cortex(Jia
350
et al., 2023). Since Jia’s work is similar to that of BMC, but with a narrower range, we
351
primarily compare and discuss our results with those from Yu. Yu identified potential
352
causal effects between decreased FA, increased MD, and ISOVF on stroke in the
353
forward MR analysis, as well as the causal effect of stroke on ISOVF in the reverse
354
MR analysis. However, these associations were not observed in our study. A
355
comparison of the two studies is presented in Supplementary Table 6. The disparity
356
in results could be attributed to two main reasons. Firstly, we utilized the largest
357
stroke GWAS results from GIGASTROKE(Mishra et al., 2022), while Yu’ s study used
358
data from MEGASTROKE(Yu et al., 2023). The utilization of a larger GWAS dataset
359
enhances the power to assess causal relationships while reducing false positives.
360
Secondly, the screening criteria for selecting IVs differed between the two studies,
361
which may have resulted in variations in the IVs chosen. We were unable to replicate
362
Yu’s tests using their IVs due to errors in their provided repetitive IV information.
363
Additionally, the significant IDP-stroke pairs identified in our study were not included
364
in Yu’s study due to their artificial selection of the IDPs before the MR study.
365
366
Our study employed bidirectional TSMR using SNPs as instrumental variables to
367
investigate the causal relationships between IDPs and stroke. This approach mitigated
368
environmental or lifestyle confounding factors and demonstrated greater effectiveness
369
compared to observational case-control studies. Additionally, we utilized the largest
370
available GWAS summary data for both exposure and outcome, enhancing statistical
371
power. Our findings revealed a potential causal association between the volume of
372
thalamus nuclei and the development of stroke, a previously unexplored relationship.
373
374
Despite the advantages, there are a few limitations of our study. The selection of traits
375
for bidirectional TSMR analysis was not based on careful consideration of existing
376
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324441doi: medRxiv preprint
12
observational studies, highlighting the need for further research to determine the exact
377
nature of these relationships. However, screening of the IDPs without selection of the
378
traits also ensures a comprehensive understanding of the causality relationships
379
between IDPs and stroke. The population under investigation in this study is limited
380
to individuals of European descent due to data availability constraints. Nevertheless,
381
the insights gained and the methodologies employed in this research can serve as a
382
valuable reference for future investigations among non-European populations.
383
Additionally, we applied a conservative multiple-testing correction, which may have
384
led to the exclusion of previously reported or potentially significant associations, but
385
it guaranteed the robustness of our results. For instance, our study did not find a
386
significant causal association between MD in the superior fronto-occipital fasciculus
387
and stroke, despite its reported significance in Yu’s work(Yu et al., 2023). This
388
discrepancy could be attributed to differences in IV screening conditions and/or the
389
utilization of a larger database. Further validation on which is the reasons for the
390
discrepancy will require a complete release of the IVs in Yu’s work. We have made
391
our IVs completely publicly available to facilitate the reproducibility of our study.
392
393
5.Conclusion
394
In summary, our study utilized bidirectional TSMR analysis with the largest available
395
GWAS summary data to investigate the causal relationship between brain IDPs and
396
stroke or its subtypes. We identified causal associations between the volume of
397
thalamus nuclei and stroke, contributing to our understanding of the link between
398
brain anatomy structure and stroke. These findings may have implications for
399
potential strategies in predicting and intervening in stroke risk at the brain imaging
400
level. By identifying individuals with specific brain IDPs that potentially lead to
401
stroke, clinicians can recommend appropriate lifestyle changes or monitor their health
402
status accordingly. This could facilitate the implementation of high-risk strategies
403
more conveniently.
404
405
Author contributions
406
The study was conceived and designed by SL, SYC. GWAS summary data was
407
collected by SYC, XCL, XYY, LSL, PYS, YXT, JNW. IV selection, MR analysis and
408
manuscript were done by XYY. Manuscript was revised by SL, SYC, XYY.
409
410
Ethics approval statement
411
The study has obtained ethical approval as documented in previous GWAS research.
412
Participants have provided informed consent prior to their participation in the study.
413
414
Conflict of interest
415
The authors declare no competing interests.
416
417
Data availability
418
The GWAS summary statistics utilized in our study were sourced from the largest
419
scale published GWA studies on both IDPs and stroke. Further information of our
420
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324441doi: medRxiv preprint
13
study are provided in Supplementary tables and Supplementary Figures. All
421
supplementary data for this study are accessible in online open-access repository
422
(https://github.com/liusylab/IDP_Stroke_2SMR).
423
424
Appendix A. Supplementary Tables
425
Further information about the data we used can be obtained in Supplementary
426
Tables.
427
Appendix B. Supplementary Figures
428
Further illustrations of our study is presented in Supplementary Figures.
429
430
Acknowledgement
431
This study is supported by grants from National Key R&D Program of China
432
(2022YFC2502400, 2022YFC2502402, 2022YFE0209600), the National Natural
433
Science Foundation of China (82101359), Young Elite Scientists Sponsorship
434
Program by CAST (2023QNRC001).
435
436
Reference:
437
Baleydier, C., and Mauguiere, F. (1980). The duality of the cingulate gyrus in monkey.
438
Neuroanatomical study and functional hypothesis. Brain 103, 525–554. doi:
439
10.1093/brain/103.3.525
440
Bezdudnaya, T., and Keller, A. (2008). Laterodorsal nucleus of the thalamus: A
441
processor of somatosensory inputs. J Comp Neurol 507, 1979–1989. doi:
442
10.1002/cne.21664
443
Bowden, J., Davey Smith, G., and Burgess, S. (2015). Mendelian randomization with
444
invalid instruments: effect estimation and bias detection through Egger
445
regression. Int J Epidemiol 44, 512–525. doi: 10.1093/ije/dyv080
446
Burgess, S., Dudbridge, F., and Thompson, S. G. (2016). Combining information on
447
multiple instrumental variables in Mendelian randomization: comparison of
448
allele score and summarized data methods. Stat Med 35, 1880–1906. doi:
449
10.1002/sim.6835
450
Campbell, B. C. V., De Silva, D. A., Macleod, M. R., Coutts, S. B., Schwamm, L. H.,
451
Davis, S. M., et al. (2019). Ischaemic stroke. Nat Rev Dis Primers 5, 70. doi:
452
10.1038/s41572-019-0118-8
453
Campbell, B. C. V., and Khatri, P. (2020). Stroke. Lancet 396, 129–142. doi:
454
10.1016/S0140-6736(20)31179-X
455
Carrera, E., Michel, P., and Bogousslavsky, J. (2004). Anteromedian, central, and
456
posterolateral infarcts of the thalamus: three variant types. Stroke 35,
457
2826–2831. doi: 10.1161/01.STR.0000147039.49252.2f
458
Casanova, C., Nordmann, J. P., and Molotchnikoff, S. (1991). [Pulvina-lateralis
459
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324441doi: medRxiv preprint
14
posterior nucleus complex of mammals and the visual function]. J Physiol
460
(Paris) 85, 44–57.
461
c-Fos expression in the limbic thalamus following thermoregulatory and wake-sleep
462
changes in the rat - PubMed (n.d.). Available at:
463
https://pubmed.ncbi.nlm.nih.gov/30887077/ (Accessed March 3, 2024).
464
Conrad, J., Habs, M., Ruehl, R. M., Bögle, R., Ertl, M., Kirsch, V., et al. (2022).
465
White matter volume loss drives cortical reshaping after thalamic infarcts.
466
Neuroimage Clin 33, 102953. doi: 10.1016/j.nicl.2022.102953
467
Feigin, V. L., Stark, B. A., Johnson, C. O., Roth, G. A., Bisignano, C., Abady, G. G.,
468
et al. (2021). Global, regional, and national burden of stroke and its risk
469
factors, 1990–2019: a systematic analysis for the Global Burden of Disease
470
Study 2019. The Lancet Neurology 20, 795–820. doi:
471
10.1016/S1474-4422(21)00252-0
472
Forno, G., Saranathan, M., Contador, J., Guillen, N., Falgàs, N., Tort-Merino, A., et al.
473
(2023). Thalamic nuclei changes in early and late onset Alzheimer’s disease.
474
Curr Res Neurobiol 4, 100084. doi: 10.1016/j.crneur.2023.100084
475
Geng, J., Gao, F., Ramirez, J., Honjo, K., Holmes, M. F., Adamo, S., et al. (2023).
476
Secondary thalamic atrophy related to brain infarction may contribute to
477
post-stroke cognitive impairment. J Stroke Cerebrovasc Dis 32, 106895. doi:
478
10.1016/j.jstrokecerebrovasdis.2022.106895
479
Gong, W., Beckmann, C. F., and Smith, S. M. (2021). Phenotype discovery from
480
population brain imaging. Med Image Anal 71, 102050. doi:
481
10.1016/j.media.2021.102050
482
Hartwig, F. P., Davey Smith, G., and Bowden, J. (2017). Robust inference in summary
483
data Mendelian randomization via the zero modal pleiotropy assumption. Int J
484
Epidemiol 46, 1985–1998. doi: 10.1093/ije/dyx102
485
Ilves, N., Lõo, S., Ilves, N., Laugesaar, R., Loorits, D., Kool, P., et al. (2022).
486
Ipsilesional volume loss of basal ganglia and thalamus is associated with poor
487
hand function after ischemic perinatal stroke. BMC Neurol 22, 23. doi:
488
10.1186/s12883-022-02550-3
489
Jia, Y., Sun, H., Sun, L., Wang, Y., Xu, Q., Liu, Y., et al. (2023). Mendelian
490
randomization analysis implicates bidirectional associations between brain
491
imaging-derived phenotypes and ischemic stroke. Cereb Cortex 33,
492
10848–10857. doi: 10.1093/cercor/bhad329
493
Kaffashian, S., Tzourio, C., Zhu, Y.-C., Mazoyer, B., and Debette, S. (2016).
494
Differential Effect of White-Matter Lesions and Covert Brain Infarcts on the
495
Risk of Ischemic Stroke and Intracerebral Hemorrhage. Stroke 47, 1923–1925.
496
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324441doi: medRxiv preprint
15
doi: 10.1161/STROKEAHA.116.012734
497
Kim, Y. S., Kim, B. J., Noh, K. C., Lee, K. M., Heo, S. H., Choi, H.-Y., et al. (2019).
498
Distal versus Proximal Middle Cerebral Artery Occlusion: Different
499
Mechanisms. Cerebrovasc Dis 47, 238–244. doi: 10.1159/000500947
500
Liew, S.-L., Zavaliangos-Petropulu, A., Schweighofer, N., Jahanshad, N., Lang, C. E.,
501
Lohse, K. R., et al. (2021). Smaller spared subcortical nuclei are associated
502
with worse post-stroke sensorimotor outcomes in 28 cohorts worldwide. Brain
503
Commun 3, fcab254. doi: 10.1093/braincomms/fcab254
504
Liu, J., Wang, C., Qin, W., Ding, H., Peng, Y., Guo, J., et al. (2023). Cortical structural
505
changes after subcortical stroke: Patterns and correlates. Hum Brain Mapp 44,
506
727–743. doi: 10.1002/hbm.26095
507
Miller, K. L., Alfaro-Almagro, F., Bangerter, N. K., Thomas, D. L., Yacoub, E., Xu, J.,
508
et al. (2016). Multimodal population brain imaging in the UK Biobank
509
prospective epidemiological study. Nat Neurosci 19, 1523–1536. doi:
510
10.1038/nn.4393
511
Mishra, A., Malik, R., Hachiya, T., Jürgenson, T., Namba, S., Posner, D. C., et al.
512
(2022). Stroke genetics informs drug discovery and risk prediction across
513
ancestries. Nature 611, 115–123. doi: 10.1038/s41586-022-05165-3
514
Motles, E., Infante, C., and González, M. (1985). [Role of the pulvinar lateralis
515
posterior complex on the rotation response. Relation with other cerebral
516
structures. Pharmacological systems involved in this behavior]. Acta Physiol
517
Pharmacol Latinoam 35, 237–249.
518
Predicted Brain Age After Stroke - PubMed (n.d.). Available at:
519
https://pubmed.ncbi.nlm.nih.gov/31920628/ (Accessed March 20, 2024).
520
Sakayori, N., Kato, S., Sugawara, M., Setogawa, S., Fukushima, H., Ishikawa, R., et
521
al. (2019). Motor skills mediated through cerebellothalamic tracts projecting
522
to the central lateral nucleus. Mol Brain 12, 13. doi:
523
10.1186/s13041-019-0431-x
524
Shibata, H., and Naito, J. (2005). Organization of anterior cingulate and frontal
525
cortical projections to the anterior and laterodorsal thalamic nuclei in the rat.
526
Brain Res 1059, 93–103. doi: 10.1016/j.brainres.2005.08.025
527
Shibata, H., and Yoshiko, H. (2015). Thalamocortical projections of the anteroventral
528
thalamic nucleus in the rabbit. J Comp Neurol 523, 726–741. doi:
529
10.1002/cne.23700
530
Siemonsen, S., Mouridsen, K., Holst, B., Ries, T., Finsterbusch, J., Thomalla, G., et al.
531
(2009). Quantitative t2 values predict time from symptom onset in acute stroke
532
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324441doi: medRxiv preprint
16
patients. Stroke 40, 1612–1616. doi: 10.1161/STROKEAHA.108.542548
533
Smith, S. M., Douaud, G., Chen, W., Hanayik, T., Alfaro-Almagro, F., Sharp, K., et al.
534
(2021). An expanded set of genome-wide association studies of brain imaging
535
phenotypes in UK Biobank. Nat Neurosci 24, 737–745. doi:
536
10.1038/s41593-021-00826-4
537
Vertes, R. P., Linley, S. B., and Rojas, A. K. P. (2022). Structural and functional
538
organization of the midline and intralaminar nuclei of the thalamus. Front
539
Behav Neurosci 16, 964644. doi: 10.3389/fnbeh.2022.964644
540
Wang, Z., Zhang, S., Liu, C., Yao, Y., Shi, J., Zhang, J., et al. (2019). A study of
541
neurite orientation dispersion and density imaging in ischemic stroke.
542
Magnetic Resonance Imaging 57, 28–33. doi: 10.1016/j.mri.2018.10.018
543
Xu, J., Galardi, M. M., Pok, B., Patel, K. K., Zhao, C. W., Andrews, J. P., et al. (2020).
544
Thalamic Stimulation Improves Postictal Cortical Arousal and Behavior. J
545
Neurosci 40, 7343–7354. doi: 10.1523/JNEUROSCI.1370-20.2020
546
Yang, J., Lee, S. H., Goddard, M. E., and Visscher, P. M. (2011). GCTA: a tool for
547
genome-wide complex trait analysis. Am J Hum Genet 88, 76–82. doi:
548
10.1016/j.ajhg.2010.11.011
549
Yeterian, E. H., and Pandya, D. N. (1988). Corticothalamic connections of paralimbic
550
regions in the rhesus monkey. J Comp Neurol 269, 130–146. doi:
551
10.1002/cne.902690111
552
Yu, K., Chen, X.-F., Guo, J., Wang, S., Huang, X.-T., Guo, Y., et al. (2023).
553
Assessment of bidirectional relationships between brain imaging-derived
554
phenotypes and stroke: a Mendelian randomization study. BMC Med 21, 271.
555
doi: 10.1186/s12916-023-02982-9
556
Zhang, Q., Wang, D., Wu, S., Ren, Y., Li, Y., Zhang, J., et al. (2021). Diffuse Tract
557
Damage Correlates With Global Cognitive Impairment in Cerebral Autosomal
558
Dominant Arteriopathy With Subcortical Infarcts and Leukoencephalopathy: A
559
Tract-Based Spatial Statistics Study. J Comput Assist Tomogr 45, 285–293. doi:
560
10.1097/RCT.0000000000001129
561
562
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324441doi: medRxiv preprint
