Alzheimer's disease (AD) is the most primary cause of dementia, which is clinically characterized by progressive cognitive decline (Scheltens et al., 2021). Given the enormous burden of AD on family nursing and public health, early identification of AD and timely management are particularly imperative (Livingston et al., 2020). On the one hand, as a transitional stage between normal aging and AD, mild cognitive impairment (MCI) is an appropriate period for early diagnosis and prevention of AD. On the other hand, around 9% of dementia cases suffer from hearing loss (HL), which has the highest population attributable fraction for dementia (Livingston et al., 2020). Consequently, understanding the association between HL and MCI may potentially provide a hint on the early management of AD. Several studies have indicated the relationship between HL and cognitive decline, and the role of HL as an increased risk for dementia (Deal et al., 2017; Thomson et al., 2017; Wei et al., 2017; Zheng et al., 2017). For example, David G. Loughrey found that age-related HL was related to increased risk of dementia (Loughrey et al., 2018). Similarly, a longitudinal study of 836 subjects found that elders with HL had an higher risk of developing dementia and a faster decline of cognition (Gurgel et al., 2014). Cognitive decline was often accompanied by the alternation in brain structure and function. Further evidence can be found in several previous morphological neuroimaging studies. For example, HL subjects suffer from microstructural dysfunction of auditory fiber tracts (Chang et al., 2004; Golub, 2017) and more rapid development of brain atrophy, especially in the temporal lobe (Eckert et al., 2012; Lin et al., 2014; Di Stadio et al., 2021). However, further research should be done to reflect the regional neuronal abnormalities.

Identification of regional neuronal abnormalities is crucial to understand the underlying mechanism behind neurodegenerative diseases. Resting-state functional magnetic resonance imaging (rsfMRI) is sensitive to imperceptible brain functional changes (Hohenfeld et al., 2018). Moreover, fractional amplitude of low-frequency fluctuation (fALFF) can effectively reflect spontaneous brain activities (Zou et al., 2008) thus has been widely applied in AD related research (Yang et al., 2020; Marin-Marin et al., 2021). For example, Liu et al. observed the fALFF in AD spectrum and found that the progression of AD could be portrayed by abnormal spontaneous regional neural activity (Yang et al., 2020). In addition, 18F-fluorodeoxyglucose positron emission tomography (FDG-PET) is a reliable method to reflect regional neuron glycometabolism (Kato et al., 2016; Caminiti et al., 2018). Blazhenets et al. demonstrated that FDG-PET had a reliable predictive value for the progression of MCI patients (Blazhenets et al., 2020). Hence, the combination of fALFF and FDG-PET could better illustrate the local brain activity in prodromal AD with HL.

In this study, we aim to explore the effect of HL on spontaneous brain activity in early stage of AD by combining the fALFF with FDG-PET. According to previous studies (Martini et al., 2014; Fulton et al., 2015; Wayne and Johnsrude, 2015; Griffiths et al., 2020), the long-term effect of HL on cognition was negative, while the early effect was unclear. We thus hypothesize that (1) the spontaneous brain activity of cognitively normal (CN) and MCI populations may respond differently to HL; (2) such functional changes may correlate with cognition.

Our study investigated the relations between HL and cognitive status reflected by spontaneous brain activity in vivo. We found a strong interactive effect between HL and cognitive status in the left STG and bilateral ITG. Further analysis showed that HL has different effects on CN and MCI. CN had increased spontaneous brain activity in auditory cortex and decreased activity in ITG. Such pattern altered with disease stage changing, demonstrated as decreased activity of the auditory cortex and increased activity in ITG in MCI. In general, HL and cognition do have an interactive effect on brain function, which may serve as the underlying mechanism of cognitive decline in the HL population. Our study may provide insights for the underlying mechanism behind HL in AD and suggest that early treatment of HL may postpone the transition of dementia.

The interactive effect (HL × cognitive status) we found mainly located in the regions involved in auditory (left STG) and visuospatial function (bilateral ITG). STG is the auditory cortex center (Tasaka et al., 2020) which receives auditory information from bilateral sources (mainly contralateral) (Brewer and Barton, 2016). Previous studies have shown that STG is mainly involved in extracting linguistic traits from speech input (Lin et al., 2014; Chern and Golub, 2019; Yi et al., 2019). In addition, previous studies found that brain atrophy, especially in STG, was associated with HL (Lin et al., 2014; Di Stadio et al., 2021). Similarly, a diffusion imaging study showed that there is a relationship between the left STG volume and auditory hallucination proneness (Spray et al., 2018). All these evidences support our findings to some extent. HL may be associated with degeneration of the auditory cortex. It is worth noting that above-mentioned results were only observed in the left STG but not the right side, which may indicate the spatial lateralization and related to the laterality of speech production (Kell et al., 2011; Hodgson and Hudson, 2018; Quass et al., 2018). This auditory lateralization was proposed by Christian Alexander Kell, who reported that speech production was a lateralized brain activity that could arise from a left dominance either in speech executive or sensory processes or both (Kell et al., 2011). Another study found that thickness of the left STG was correlated with cognitive performance, executive function, information processing speed, and attention (Achiron et al., 2013). Moreover, we also found a significant interactive effect (HL × cognitive status) on ITG, which is generally thought to be used to process visual information (Onitsuka et al., 2004) and multimodal sensory integration (Onitsuka et al., 2004). Some studies have also suggested that ITG may be involved in the integration of auditory information (Alain et al., 2018; Manno et al., 2021).

The cognitive load hypothesis may explain the functional changes that occur in these two brain regions. This hypothesis theorizes that HL may lead to the degradation of auditory signals, greater cognitive resources being required for auditory perceptual processing, and transfer from other cognitive tasks to effortful listening, ultimately resulting in cognitive reserve depletion (Fulton et al., 2015; Uchida et al., 2019). STG and ITG may be involved in auditory reception (Adank, 2012; Alain et al., 2018), processing (Du et al., 2016; Eckert et al., 2016), and integration (Onitsuka et al., 2004). In patients with HL, these regions may work together to keep auditory perception and cognition performance, and working patterns may change with the severity of the disease. Our results demonstrated a distinct effect pattern of HL on CN and MCI.

As for CN, CN+HL showed decreased fALFF in bilateral ITG, and an increasing trend of fALFF in STG compared to CN+NHL, indicating impaired function of information input. HL may lead to decline in speech input (Peelle et al., 2011; Eckert et al., 2012). CN subjects with HL may compensate the ability to process auditory information so as to maintain a relatively normal hearing experience (Pichora-Fuller et al., 1995; Tun et al., 2009). Nevertheless, compensation mechanism requires occupying global cognitive resources, which may lead to decreased activity in other brain regions (Tun et al., 2009; Martini et al., 2014). As a result, decreased activity in these occupied brain regions may lead to a decline in related cognitive performance, such as vision, attention, and working memory (Tun et al., 2009; Lin and Albert, 2014; Wayne and Johnsrude, 2015). The results also support the idea that the increased activity in local brain regions is actually the result of compensation. Our results showed that the fALFF value in the left STG was negatively correlated with MMSE, and the fALFF value in the right ITG was negatively correlated with visuospatial function. Since STG and ITG are spatially close to each other and both play a role in auditory processing and integration (Onitsuka et al., 2004), STG activity increased to compensate for the loss while ITG activity decreases when HL is present (Martini et al., 2014; Gagné et al., 2017). According to a study, the auditory processing regions, such as STG, are recruited by the visual system (ITG) when auditory input lost due to auditory deprivation (Manno et al., 2021). As mentioned above, the cognitive load theory may be the underlying mechanism between HL in MCI. Nevertheless, such pattern may change along AD stage according to our result.

As for MCI stage, MCI+HL showed increased fALFF in ITG and decreased fALFF in STG compared to CN+HL. This suggested decreased neuronal activity in STG of the MCI population compared to the CN population, while the neuronal activity in ITG is increased when HL is present. Such result could be explained by the effect of compensation. Generally speaking, long-term HL is always related to degeneration of the auditory cortex neuron (Chatani et al., 2016). Aforementioned compensatory effects can partly alleviate the effects of HL, but it might aggravate the burden on neurons accordingly, and the burden may start early and causing irreversible damage to neurons (Boyle et al., 2008). This explains the lower STG activity of MCI population compared to CN population. Despite this increased activity, widespread brain regions are activated during compensation to improve hearing experience (Pichora-Fuller et al., 1995; Griffiths et al., 2020). Studies have shown that compared to younger people, healthy older people requires broader brain activation to understand sentences of the same difficulty (Wingfield and Grossman, 2006; Vaden et al., 2015). This suggested that older people may have poorer auditory cortex function than younger people due to aging thus requires more cognitive resources for auditory information processing (Wingfield and Grossman, 2006).

As expected, we found that the fALFF value was adversely correlated with the FDG SUVR value in the left ITG. Although local neuronal activity is enhanced, some neurons in the same regions have been damaged at the same time due to overload. Studies have shown that in the case of long-term over-occupation, the function of related brain regions might be affected (Wingfield and Grossman, 2006). In addition, we also found the FDG SUVR value in STG was positively correlated with MoCA and language, and the FDG SUVR value in ITG was positively correlated with MMSE, executive function. As many studies have shown, HL is associated with cognitive decline, including but not limited to, general cognition, executive function and memory (Campbell and Sharma, 2013; Rönnberg et al., 2013; Gurgel et al., 2014).

There are several limitations to this study. Firstly, our sample size was relatively small. Due to the limited sample size, the current results should be treated with caution. Secondly, because the study was a cross-sectional design, the causal relationship between HL and cognition decline is unclear yet. But our results also provide novel insights into the mechanism between HL and cognition decline. It is necessary to repeat the investigation in a larger and longitudinal sample in the future. Thirdly, most participants with HL were male. This may be due to the gender distribution of HL. Some studies have shown that there is a gender difference in the incidence of hearing loss, such as higher incidence (Mitchell et al., 2011; Jafari et al., 2019) and earlier onset in males (Pearson et al., 1995). Nevertheless, gender has been adjusted as a covariate in current study. It is necessary to conduct research in the female population in future. Finally, the information of comorbidities (e.g., tinnitus and hyperacusis) and auditory amplification was unavailable. Thus, we were unable to examine whether comorbidities or auditory amplification affected the association between HL and neurodegeneration. Since this is only an exploratory study at present, future studies will consider these influencing factors and conduct longitudinal studies to investigate the impact of HL in brain activity along the AD continuum, from normal aging to more severe clinical stages.

Our study demonstrated that the interaction between HL and cognition could be reflected by spontaneous brain activity (STG and ITG). As mentioned above, HL may have different effects on NC and MCI stages. Cognitive load theory may be the underlying mechanism behind this. This compensatory effect might also be responsible for cognitive decline in people with HL. This work may provide insights for the underlying mechanism behind HL and AD, suggesting that early treatment of HL may be meaningful to prevent AD.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: http://adni.loni.usc.edu/.

All procedures performed in studies involving human participants were under the ethical standards of the Institutional and National Research Committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Written informed consent was from all participants and authorized representatives, and the study partners before any protocol-specific procedures were carried out in the ADNI study. More details from http://www.adni-info.org.

LH, QZ, and XLuo designed the study and wrote the first draft of the manuscript. LH and KL analyzed the MRI data and wrote the protocol. LH and QZ collected clinical and MRI data. XX, XLiu, ZLi, YF, YW, TZ, YC, ZLiu, PH, and MZ assisted with research design and interpretation of results. All authors contributed to the final manuscript, read, and approved the final manuscript.

This study was funded by National Key Research and Development Program of China (Grant No. 2016YFC1306600), National Natural Science Foundation of China (Grant Nos. 81901707, 82001766, and 81701337), the Fundamental Research Funds for the Central Universities (No. 2017XZZX001-01), and Zhejiang Medicine and Health Science and Technology Program (2018KY418). The data collection and sharing for this project were funded by the ADNI (National Institutes of Health Grant U01 AG024904) and DOD ADNI (Department of Defense Award No. W81XWH-12-2-0012). ADNI was funded by the NIA, the NIBIB, and through generous contributions from the following: AbbVie, Alzheimer's Association; Alzheimer's Drug Discovery Foundation; Bio Clinica, Araclon Biotech; Inc.; Biogen; Bristol-Myers Squibb Company; Cere Spir, Inc.; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; F. Hoffmann-La Roche Ltd., Euro Immun; and its affiliated company Genentech, Inc.; Fujirebio; IXICO Ltd.; GE Healthcare; Johnson & Johnson Pharmaceutical Research and Development LLC.; Janssen Alzheimer Immunotherapy Research and Development, LLC.; Lumosity; Lundbeck; Merck & Co., Inc.; MesoScale Diagnostics, LLC.; Neurotrack Technologies; NeuroRx Research; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Servier; Piramal Imaging; Transition Therapeutics; and Takeda Pharmaceutical Company. The Canadian Institutes of Health Research provides funds to support ADNI clinical sites. Private sector contributions are facilitated by the Foundation for the National Institutes of Health (www.fnih.org). The grantee organization is the Northern California Institute for Research and Education, and the study is coordinated by the AD Cooperative Study at the University of California, San Diego.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.