Bayesian lesion-deficit inference with Bayes factor mapping: Key advantages, limitations, and a toolbox

Sperber, Christoph; Gallucci, Laura; Smaczny, Stefan; Umarova, Roza

doi:10.1016/j.neuroimage.2023.120008

Cited by 10 publications

(31 citation statements)

References 64 publications

(100 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For basal ganglia stroke, the Bayesian analysis (Figure 3A) was, as expected 50 , more liberal and found scattered clusters of evidence for lesion network-neglect associations (i.e.…”

Section: Resultssupporting

confidence: 68%

“…Second, we replicated the analysis with a Bayesian approach with Bayes factor mapping to gain insights into evidence for the null hypothesis h 0 and statistical power. We used the Bayesian Lesion-Deficit Inference toolbox 50 with a design equivalent to the frequentist analysis. Bayesian general linear models tested if an association between cortical lesion-network connectivity and spatial neglect exists in two-sided tests against the null hypothesis that no such association exists.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Spatial neglect after subcortical stroke: sometimes a cortico-cortical disconnection syndrome

Sperber,

Rosenzopf,

Wawrzyniak

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

Background and Objectives: Spatial neglect is commonly attributed to lesions of a predominantly right-hemispheric cortical network. Although spatial neglect was also repeatedly observed after lesions to the basal ganglia and the thalamus, many anatomical network models omit these structures. We investigated if disruption of functional or structural connectivity can explain spatial neglect in subcortical stroke. Methods: We retrospectively investigated data of first-ever, acute stroke patients with right sided lesions of the basal ganglia (n = 27) or the thalamus (n = 16). Based on lesion location, we estimated i) functional connectivity via lesion-network mapping with normative resting state fMRI data, ii) structural white matter disconnection and iii) tract-wise disconnection of association fibres based on normative tractography data to investigate the association of spatial neglect and disconnection measures. Results: Apart from very small clusters of functional disconnection observed in inferior/middle frontal regions in lesion-network symptom mapping for basal ganglia lesions, our analyses found no evidence of functional or structural subcortico-cortical disconnection. Instead, the multivariate consideration of lesion load to several association fibres predicted the occurrence of spatial neglect (p = 0.0048; AUC = 0.76), which were the superior longitudinal fasciculus, inferior occipitofrontal fasciculus, superior occipitofrontal fasciculus, and the uncinate fasciculus. Conclusion: Disconnection of long (cortico-cortical) association fibres can explain spatial neglect in subcortical stroke. Like the competing theory of remote cortical hypoperfusion, this mechanism does not require the assumption of a genuine role for subcortical grey matter structures in spatial neglect.

show abstract

“…For basal ganglia stroke, the Bayesian analysis (Figure 3A) was, as expected 50 , more liberal and found scattered clusters of evidence for lesion network-neglect associations (i.e.…”

Section: Resultssupporting

confidence: 68%

Section: Discussionmentioning

confidence: 99%

Spatial neglect after subcortical stroke: sometimes a cortico-cortical disconnection syndrome

Sperber,

Rosenzopf,

Wawrzyniak

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…BLDI is statistically very liberal and susceptible to overestimating the neural anatomy associated with a deficit (Sperber et al, 2023). Thus, for better comparability of the Bayesian and the multivariate analyses, we focused our interpretation of the results on the peak Bayes factors (see Figure 3), at a threshold similar to that of the multivariate analysis which included a correction for multiple comparisons.…”

Section: Bayesian Lesion-deficit Inferencementioning

confidence: 99%

“…Importantly, BLDI performs better when statistical power is low (Sperber et al, 2023). As such, this approach may be better suited to detect functional contributions in areas of the brain less frequently damaged by stroke lesions, such as, for example, the superior parietal lobe and IPS, allowing us, again, to obtain a fuller picture of the areas of the brain associated with the cognitive function of interest.…”

mentioning

confidence: 99%

The neuroanatomy of visual extinction following right hemisphere brain damage: Insights from multivariate and Bayesian lesion analyses in acute stroke

Sperber,

Wiesen,

Karnath

et al. 2024

Human Brain Mapping

Self Cite

View full text Add to dashboard Cite

Multi‐target attention, that is, the ability to attend and respond to multiple visual targets presented simultaneously on the horizontal meridian across both visual fields, is essential for everyday real‐world behaviour. Given the close link between the neuropsychological deficit of extinction and attentional limits in healthy subjects, investigating the anatomy that underlies extinction is uniquely capable of providing important insights concerning the anatomy critical for normal multi‐target attention. Previous studies into the brain areas critical for multi‐target attention and its failure in extinction patients have, however, produced heterogeneous results. In the current study, we used multivariate and Bayesian lesion analysis approaches to investigate the anatomical substrate of visual extinction in a large sample of 108 acute right hemisphere stroke patients. The use of acute stroke patient data and multivariate/Bayesian lesion analysis approaches allowed us to address limitations associated with previous studies and so obtain a more complete picture of the functional network associated with visual extinction. Our results demonstrate that the right temporo‐parietal junction (TPJ) is critically associated with visual extinction. The Bayesian lesion analysis additionally implicated the right intraparietal sulcus (IPS), in line with the results of studies in neurologically healthy participants that highlighted the IPS as the area critical for multi‐target attention. Our findings resolve the seemingly conflicting previous findings, and emphasise the urgent need for further research to clarify the precise cognitive role of the right TPJ in multi‐target attention and its failure in extinction patients.

show abstract

“…In stroke lesion-deficit mapping, control for lesion size or secondary deficits can create spurious associations. 12,13 The literature on causal inference cannot only provide a theoretical background as to when and why to apply statistical control, but it also shows awareness of the potential dangers of ill-applied covariate control. [14][15][16] When we want to infer causal effects from observational data, the strategy to identify a covariate requires making assumptions about the causal relationships between all variables.…”

Section: Introductionmentioning

confidence: 99%

Control without cause: How covariate control biases our insights into brain architecture and pathology

Sperber,

Gallucci,

Arnold

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

Inferential analysis of normal or pathological brain imaging data - as in brain mapping or the identification of neurological imaging markers - is often controlled for secondary variables. However, a rationale for covariate control is rarely given and formal criteria to identify appropriate covariates in such complex data are lacking. We investigated the impact and adequacy of covariate control in large-scale imaging data using the example of stroke lesion-deficit mapping. In 183 stroke patients, we evaluated control for age, sex, hypertension, or lesion volume when mapping real or simulated deficits. We found that the impact of covariate control varies and can be strong, but it does not necessarily improve the precision of results. Instead, it systematically shifts results towards the inversed associations between imaging features and the covariate. This effect of covariate control can bias results and, as shown in another experiment, can even create effects out of nothing. The widespread use of covariate control in the statistical analysis of clinical brain imaging data - and, likely, other biological high-dimensional data as well - may not generally improve statistical results, but it may just change them. Therefore, covariate control constitutes a problematic degree of freedom in the analysis of brain imaging data and may often not be justified at all.

show abstract

Bayesian lesion-deficit inference with Bayes factor mapping: Key advantages, limitations, and a toolbox

Cited by 10 publications

References 64 publications

Spatial neglect after subcortical stroke: sometimes a cortico-cortical disconnection syndrome

Spatial neglect after subcortical stroke: sometimes a cortico-cortical disconnection syndrome

The neuroanatomy of visual extinction following right hemisphere brain damage: Insights from multivariate and Bayesian lesion analyses in acute stroke

Control without cause: How covariate control biases our insights into brain architecture and pathology

Contact Info

Product

Resources

About