With the aging population, dementia has become a serious global public health problem. Recent estimates suggest that there are presently approximately 50 million individuals living with dementia globally, a number that is expected to triple to 152 million by 2050 (Lynch, 2020). Dementia imposes a heavy burden on individuals, families, and even society. In fact, there is a long preclinical phase before dementia, which includes cognitive decline and mild cognitive impairment (MCI), both of which carry a high risk of progressing to dementia (Calvin et al., 2019; Jessen et al., 2014; Petersen et al., 2018). Given the absence of effective drugs for dementia to date, early screening and intervention in individuals with cognitive decline may become a promising prevention strategy.

Thyroid hormones are critical for maintaining cognitive function in adults, and numerous studies have investigated the relationship between thyroid function and cognitive function in this population (Khaleghzadeh-Ahangar et al., 2022; Ritchie and Yeap, 2015). Previous studies have indicated that overt thyroid dysfunction and subclinical thyroid dysfunction are associated with cognitive decline or dementia (Aubert et al., 2017; Ma et al., 2023; Pasqualetti et al., 2015; Wieland et al., 2022). In recent years, the relationship between changes in thyroid function within the reference range and cognitive function has attracted extensive attention from researchers. Epidemiological studies have shown a positive (Beydoun et al., 2015; Forti et al., 2012; Winkler et al., 2016), negative (Cappola et al., 2015; Chaker et al., 2016; Moon et al., 2014; Santhanam et al., 2022; Szlejf et al., 2018; van Osch et al., 2004), or zero (Booth et al., 2013; Castellano et al., 2013; Hogervorst et al., 2008; Samuels et al., 2016; Tan et al., 2008; Wu et al., 2016) correlation between thyroid stimulating hormone (TSH) levels within the reference range and cognitive decline or dementia. Scholars have observed that free thyroxine (FT4) in the reference range has a positive (Hogervorst et al., 2008; Yeap et al., 2012), negative (Beydoun et al., 2013; Volpato et al., 2002), or zero correlation (Booth et al., 2013; Castellano et al., 2013; Samuels et al., 2016) with cognitive decline or dementia. Notably, the above evidence on the relationship between thyroid function within the reference range and cognitive function is derived from observational studies. However, observational research is limited by confounding, reverse causation, selection bias, and information bias, which can lead to spurious associations or incorrect inferences of causality (Sekula et al., 2016).

Mendelian randomization (MR) analysis, a novel epidemiological study strategy, can overcome the limitations of observational studies by using genetic variants as instrumental variables, offering more robust evidence for a causal relationship between an exposure and an outcome (Richmond and Davey Smith, 2022; Smith and Ebrahim, 2003). Recent MR studies have reported a causal relationship between thyroid function within the reference range and Alzheimer's disease (AD) (Li et al., 2021; Marouli et al., 2021). However, little is known about the causal relationship between thyroid function within the reference range and cognitive function. Therefore, we aimed to perform a two-sample MR analysis using publicly available genome-wide association study (GWAS) statistics to explore the impact of genetically predicted thyroid function variations within the reference range on cognitive function.