Background

It is commonly known that hearing loss is prevalent in older adults and sensorineural hearing loss (SNHL) is a widespread occurrence on a global scale [62]. The prevalence of hearing loss doubles with each successive decade of increased age and mostly begins during mid-life. For instance, approximately 27% of adults in their mid-life experience hearing loss [15]. However, there is a scarcity of research highlighting the increasing prevalence of hearing loss in mid-life. Such findings present potential opportunities for early intervention to mitigate the significant future impact of age-related hearing loss [84]. Hearing loss in mid-life not only acts as a potential risk factor for dementia [23] in older adults by influencing changes in brain structures but is also associated with increased late-life atrophy over time, particularly in specific temporal lobe structures involved in auditory processing and AD [6]. In recent studies, a strong link has been established between dementia and hearing loss, with the latter being identified as an independent but also a promising modifiable risk factor for incident dementia [3, 19, 20, 31,32,33, 37, 39, 56, 57, 69, 72, 81]. The causal mechanisms are discussed in detail elsewhere [86].

To date, there is limited data [63] that attempted to identify subtle cognitive changes that occur in early mid-life adults (between 45 and 55 years of age) with hearing loss as it has been observed that 55 years is the youngest mean age [32] at which the presence of hearing loss increased the risk of dementia. Previous studies have considered a wide array of neuropsychological tests to assess the cognitive abilities of their study participants including tests to examine the executive function (EF) [61]. There are three core components of EF, namely, cognitive flexibility, inhibitory control, and working memory [65]. The Trail Making Test (TMT) [71] is a test that employs these components for its execution. Part A of the TMT primarily evaluates attention, visual scanning, speed of eye-hand coordination, and information processing whereas Part B assesses working memory and EF, specifically the capacity to transition between sets of stimuli [54, 64, 73]. Previous studies have demonstrated that TMT can effectively predict changes in individuals with MCI [14, 26, 38] and AD [16]. The TMT has been sensitive to EF [73] impairments and has consistently produced significant findings across multiple clinical [8, 36, 51, 67] and aging population [60]. Moreover, its sensitivity also extends to various neurological processes [53, 64, 76]. Although completion times for TMT-A and B are regularly used in clinical practice, derived scores are also employed for interpretive purposes [54, 64, 78]. Another noteworthy measure of interest is the number of performance errors [4, 8, 79]. The most commonly utilized derived scores include the difference score (B-A), the ratio score (B/A), and the proportion score (B-A/A). The rationale for employing these scores is twofold. Firstly, they serve to mitigate the influence of individual variability in Part B by utilizing each subject as their own control. Secondly, these scores help eliminate the contributions of motor speed and visual scanning speed by comparing Part B to Part A. Several authors have contended that these derived scores represent more refined measures of EF compared to the direct scores [5, 9, 18, 24, 41,42,43, 50, 73, 79]. While the derived scores are recognized for their sensitivity to frontal lobe damage [79], it is worth noting that they emphasize distinct characteristics. The difference score is found to be influenced by demographic characteristics when compared to the ratio score [24, 42, 50]. Conversely, the sum and multiplication scores offer indicators of overall cognitive performance rather than focusing solely on EF [43, 51]. The present study aimed to compare the performance of the TMT direct scores (i.e., completion time and performance errors for Part A and Part B) and derived scores (i.e., difference, ratio, proportion, sum, and multiplication scores) in mid-life adults with and without hearing loss. It was hypothesized a priori that adults with hearing loss would perform similarly to those without hearing loss on the TMT.

Methods

Study participants

The sample size for the present study was calculated using G*power analyses [27] utilizing the mean and standard deviation obtained from the pilot study at the significance level (α) of 0.05 and power of the test (1-ꞵ) at 0.80 [75]. The analysis prescribed a sample size ranging from 5 to 10. Due to the availability of participants, a total of 50 participants were recruited for the study consisting of 25 adults with bilateral unaided mild hearing loss (AHL) (13 female,mean age: 50.44 ± 2.84 years) and 25 adults with normal hearing (ANH) (13 female,mean age: 49 ± 3.29 years). A purposive sampling method was employed to collect samples for the present study. Demographic information for both groups is reported in Table 1.

Table 1 Demographic characteristics of participants
Full size table

All participants were right-handed, native Kannada speakers, with normal or corrected vision. None of them had a history of communication disorders, learning disabilities, neurological disorders, traumatic brain injury, or psychiatric disorders. Individuals with a history of substance abuse, use of psychoactive medications, known causes of hearing loss such as noise exposure, injuries, ototoxicity, unilateral or bilateral continuous tinnitus, those using hearing aids and experiencing significant vision issues such as untreated cataracts, macular degeneration, glaucoma, and retinopathy were excluded. All the participants had more than 12 years of formal education and underwent the Montreal Cognitive Assessment (MoCA), with a cut-off score of < 26 to eliminate potential cognitive impairment. Information regarding the presence or absence of hypertension, diabetes, smoking, alcohol consumption, and regular physical activity was also noted as these factors play an important role in accelerating cognitive decline.

Ethical considerations

The present study was approved by the Ethical Committee for bio-behavioral research involving human subjects at the All India Institute of Speech and Hearing (AIISH), Mysuru, India (Reference code: No.DOR.9.1/Ph.D/PC/930/2021-22). The participants were recruited for the study only after obtaining their written consent as per the ethical guidelines for bio-behavioral research involving human subjects of the All India Institute of Speech and Hearing [11]. All the participants and their caregivers were explained about the procedure and the approximate duration required for the tests. They were assured of safety during testing and confidentiality regarding their personal details.

Audiological evaluation

All participants underwent a comprehensive audiological assessment that included an otoscopic examination to eliminate the possibility of outer ear diseases, tympanometry to confirm the normal functioning of the middle ear, reflexometry to identify any abnormalities in the reflex pathways, along with both pure-tone and speech audiometry tests. Air conduction (AC) and bone conduction (BC) pure-tone thresholds were assessed in both ears using the modified Hughson-Westlake procedure [13]. Thresholds were recorded for octave frequencies ranging from 250 Hz to 8000 Hz for AC and 250 Hz to 4000 Hz for BC, respectively. The hearing thresholds for AC and BC were obtained using TDH-50P supra-aural headphones (Telephonics, Farmingdale, NY, USA) and a B71 bone vibrator (Radioear, KIMMETRICS, Smithsburg, MD, USA), respectively. The pure tone audiometry was carried out using Inventis Piano, a two-channel diagnostic audiometer (Inventis, 35127 Padova, Italy).

Speech recognition threshold (SRT) was obtained through live presentation of a standardized spondee word list in Kannada [70] which was within ± 12 dB HL of the pure-tone average threshold. Speech identification scores in quiet were determined using standard phonemically balanced word lists in the Kannada language [89]developed at the All India Institute of Speech and Hearing. All participants achieved speech identification scores exceeding 90% in quiet conditions. Tympanograms were obtained using a 226-Hz probe tone. Acoustic reflex thresholds were assessed for both ipsilateral and contralateral responses at frequencies of 500 Hz and 1000 Hz. It was confirmed that all participants exhibited normal middle ear function in both ears, as evidenced by ‘A’ type tympanograms and normal ipsilateral and contralateral acoustic reflex thresholds at 500 Hz and 1000 Hz. The immittance evaluations were performed using GSI- Tympstar Middle ear Analyzer (Grason Stadler Inc.-GSI-61,Milford, NH, USA). All tests were conducted in acoustically treated rooms.

The hearing sensitivity of the ANH group was within normal limits of ≤ 15 dB HL at frequencies 250 Hz to 8000 Hz for AC and 250 Hz to 4000 Hz for BC. The hearing sensitivity of the AHL group was between 25 dB HL to 40 dB HL [17] for similar frequencies for both AC and BC which was diagnosed as bilateral mild SNHL by a certified Audiologist.

Since the age difference equivalent to the cognitive decline linked with a 25-dB rise in hearing loss is 7 years, [55] adults with unaided mild SNHL were considered for the study. The study was carried out during the initial visit of the participants where they were diagnosed as having hearing loss for the first time. All the participants with reduced hearing sensitivity were directed towards hearing rehabilitation which mostly suggested the use of hearing aids.

EF tests

Trail making test

Part A

Participants were instructed to connect 25 numbered circles sequentially by drawing lines connecting them as quickly and correctly as possible. Timing commenced as soon as participants were prompted to begin using a stopwatch. The examiner pressed the stopwatch as quickly and accurately as possible as soon as the participant started and finished the test. Whenever an error was made the examiner crossed out the erroneous line and provided immediate feedback. The test was discontinued if the participant failed to complete the test within 300 seconds.

Part B

Participants were instructed to connect encircled numbers and letters, alternating between the two sequences (e.g., 1-A-2-B, and so on) consecutively in a progressive manner, up to the number 13. Commencing the timing, providing feedback for errors, and discontinuing the test were conducted following the same procedures as those employed in Part A. The entire TMT test including both Part A and B was administered in accordance with the protocol outlined by Bowie and Harvey [12]. Both tests were carried out using paper and pen in a single session. Instructions were provided at the most comfortable level (MCL) through the audiometer and the participants were required to repeat the instructions before starting the test to ensure they understood the test instructions. If the participants were able to repeat the instructions appropriately, the examiner began the test. If they failed to, instructions were given to the participants again. The test was commenced only after the participants repeated the test instructions correctly. We adhered to this protocol with the concern towards participants with hearing loss as they took part in the study in an unaided condition, and our goal was to confirm that the reduction in hearing sensitivity did not impede their comprehension of the task instructions.

Scoring

The TMT yielded a total of four direct scores and five derived scores. The direct measures of performance encompassed the time taken to complete both Part A and B, as well as the number of errors made during these tests separately. From the direct scores, five derived scores were computed that included difference score (B-A), ratio score (B/A), proportion score (B-A)/A, sum score (A+B), and multiplication score (AxB/100). It is important to note that lower direct scores and higher derived scores indicate superior performance on the test.

Statistical analysis

Data analysis of this study was performed using the Statistical Package for Social Science (SPSS) version 20. Demographic data and the TMT scores are presented using descriptive statistics including the mean, standard deviation, sum of ranks, and mean ranks. Following data analysis for normality of distribution by the Shapiro–Wilk test, it was found that the TMT scores were not normally distributed. Since the scores were not normally distributed, alternative non-parametric tests have been employed in the present study. Mann-Whitney U test was carried out to find out the significance of the difference between AHL and ANH groups on the selected dependent variables that included both direct and derived TMT scores.

Results

Significant differences were observed between AHL and ANH groups for all the audiological measures as indicated by the Mann-Whitney U test in Table 2. The smaller U values obtained for the differences in values of audiological measures were all found to be significant at the .001 level. The mean ranks obtained for AHL and ANH groups were 38 and 13 respectively for right PTA (p = .001), the mean ranks obtained for left PTA for AHL and ANH groups were 38 and 13 respectively with a significance level of .001, and the mean ranks obtained for the right SRT were 37.98 and 13.02 respectively (p = .001), the mean ranks obtained for left SRT measure for AHL and ANH groups were 38 and 13 respectively with the significance level of .001. In all of the above-mentioned audiological measures, the AHL group had significantly higher values than the ANH group. However, in the case, the mean ranks obtained for right SIS percent were 16.10 and 34.90 with a significance level of .001 as well as in the case mean ranks obtained for left SIS percent were 13.72 and 37.28 with a significance level of .001, whereas the ANH group had higher values than AHL.

Table 2 Results of audiological measures between AHL and ANH groups
Full size table

Mann-Whitney U test revealed significant differences between ANH and AHL groups (refer Table 3) in most of the TMT measures except for B/A (p = .377), and B-A/A (p = .377). In the case of TMT A, the mean ranks obtained for 13.96 and 37.04 with a significance level of .001, the mean ranks obtained for TMT A error were 22.32 and 28.68 respectively with a significance level of .038, the mean rank obtained for TMT B were 14.60 and 36.40 respectively with the significance level of .025, the mean ranks obtained for TMT B error were 21.10 and 29.60 respectively with the significance level of .025. The mean ranks obtained for ANH and AHL groups for B-A parameter were 19.16 and 31.84 with a significance level of .002. The mean ranks obtained for A+B parameters were 13.52 and 37.48 with a significance level of .001. Lastly, the mean ranks obtained for AxB/100 measure were 13.32 and 37.68 respectively with the significance level of .001. In all the above-mentioned measures AHL group had higher values than the ANH group. Box plots for direct and derived measures of TMT are depicted in Figs. 1 and 2 respectively.

Table 3 Results of Mann-Whitney U test for TMT scores between AHL and ANH groups
Full size table
Fig. 1
figure 1

Box plots for direct measures of TMT for ANH and AHL groups

Full size image
Fig. 2
figure 2

Box plots for derived measures of TMT for ANH and AHL groups

Full size image

Discussion

This study examined differences in performance on direct and derived measures of TMT between AHL and ANH groups. We found significant differences between the two groups on PTA, SRT, and SIS scores. The AHL group exhibited higher scores on PTA and SRT, and lower SIS scores relative to the ANH group, indicating poorer peripheral hearing. Findings revealed significant group differences in all the direct and derived measures of TMT except the ratio and proportion scores.

The findings of the present study are in agreement with the previous studies employing TMT B, where group differences were observed between older adults with hearing loss and normal hearing controls [1, 7, 40, 56, 57, 75]. The control group performed significantly better than the clinical group in the raw completion times, number of errors, and derived scores that resulted from difference, sum, and multiplication. However, we did not find any significant difference in the ratio and proportion scores. This variability introduces the challenge of interpreting TMT performance, especially in relation to the derived scores which could be attributed to the subject’s demographic characteristics. The ratio and proportion scores are considered to be the most sensitive indices within the TMT for identifying deficits in EF. This is due to their lower susceptibility to the influence of demographic characteristics, as indicated by previous studies [18, 50]. While we did observe a notable distinction in TMT measures, we did not identify a variance in the most sensitive EF indices. This suggests the effect of demographic characteristics play an important role in disrupting the overall cognition that does not exclusively pertain to EF.

In this light, we observed the effect of hypertension, diabetes, smoking, alcohol consumption, and physical activity on TMT scores. However, our analysis did not yield any statistically significant findings. Since these conditions are previously known to be associated with increased dementia risk [59], the present results are contrary to the previous studies. One possible explanation for our results is that prior studies have primarily focused on considerably older populations compared to the participants in our study. Additionally, the long-standing effects of these conditions could potentially exert a detrimental influence on cognitive abilities over time [45].

The present study adds to the extensive research over decades that has indicated a link between sensory abilities, particularly hearing and cognitive performance [10, 44, 49, 58, 68, 80, 88]. This association suggests that cognitive decline due to hearing loss may commence during midlife, emphasizing its pivotal role as a significant modifiable risk factor for the subsequent development of dementia. While the acceptance of hearing loss as an independent and modifiable risk factor for dementia is widespread, its specific impact in midlife as a potential harbinger of cognitive consequences in later life remains inadequately elucidated [23]. An alternative perspective suggested that midlife hearing loss could potentially be an outcome of preclinical neurodegeneration rather than a causal factor [48, 85]. The present study supports this notion by demonstrating significant differences in the TMT measures between the ANH and AHL groups.

One of the significant reasons for the results obtained in the present study could potentially be attributed to methodological disparities in the previous studies, particularly in the context of audiological assessments. For instance, certain studies relied on self-assessed hearing measurements from participants, and the criteria for hearing loss varied. In some studies, the range of normal hearing was considered up to 25 dB [55]. In our study, we meticulously adhered to a rigorous and standardized methodology. Notably, at the time of our investigation, the participants had not undergone any form of auditory rehabilitation, such as the use of hearing aids. In contrast, the participant groups in prior studies displayed heterogeneity, with some individuals employing hearing aids while others did not, considering that hearing loss is a reversible sensory deficit [66]. By diligently controlling for these variables, our study potentially unveiled significant differences across various TMT measures, shedding light on the impact of these methodological distinctions.

The observed variation in TMT performance could be due to alterations in auditory pathways resulting from peripheral hearing loss, which in turn may affect cognitive control [29]. Cohort studies suggest that even a mild degree of hearing loss may increase the risk of cognitive decline in individuals who initially exhibit cognitive intactness but are hearing-impaired at the baseline [21, 22, 30, 32, 39, 55,56,57, 83, 87]. The literature has put forth several potential hypotheses to elucidate this connection between hearing loss and cognitive changes including the cognitive load on perception hypothesis, the information degradation hypothesis, the sensory deprivation hypothesis, and the common cause hypothesis [82]. These mechanisms have the potential to deplete cognitive resources that are otherwise available for various cognitive functions. This, in turn, could contribute to the depletion of cognitive reserve and hasten the emergence of preclinical indicators of cognitive decline leading to dementia [2, 77].

On a similar line, a study among Chinese centenarians and oldest-old adults revealed that the occurrence of cognitive decline and depression was notably higher among individuals with HL compared to those without. Participants with HL exhibited significantly lower cognitive test scores, suggesting more pronounced cognitive decline, and experienced higher levels of depression [28]. Further, a systematic review of twenty-three studies found a substantial link between HL and mild cognitive impairment (MCI) [52]. The majority of studies reviewed indicated a notable association between peripheral HL and MCI. This finding is significant as EF is one of the areas commonly impacted in individuals with MCI which again supports the link between HL and subsequent cognitive decline. Research focusing on performance within distinct cognitive areas, such as processing speed, episodic memory, and EF, rather than overall cognitive assessments, has revealed moderate connections with hearing loss across these domains [40]. Numerous cohort studies and meta-analyses have highlighted a correlation between SNHL and cognitive impairment. Specifically, using pure tone audiometry, HL has been linked to lower scores across various neuropsychological tests evaluating multiple cognitive domains, including memory, language, and EF [74]. This association has been consistently observed across both small and large observational studies [25, 34, 46], as well as in cross-sectional and longitudinal population-based studies. Recent studies investigating the link between hearing loss and cognitive decline have utilized EF tests including TMT, indicating its effectiveness in this context [31]. A recent study demonstrated that steady assessments of fluid reasoning which is a part of EF and pure-tone auditory function showed considerable alteration over about seven years. During this timeframe, there was a decline in metabolic health and cognitive abilities, alongside an increase in hearing thresholds. These results suggest that significant age-related changes occur across various functional domains, even among middle-aged who are relatively healthy. Furthermore, the study found that lower cognitive performance correlated with elevated audiometric thresholds [35]. Another study that reviewed structural MRI studies revealed that HL is associated with accelerated atrophy of both total and regional brain volumes, along with reduced white matter integrity. Additionally, resting-state and task-based functional MRI studies indicated alterations in spontaneous neural activity and brain functional connectivity. It was concluded that these changes affect brain regions involved in auditory, language, cognitive, and affective processing independently of age [47]. From this perspective, it is clearly evident that HL impacts cognitive performance through a series of physiological changes that ultimately hinder brain function.

Limitations and future directions

To the best of our knowledge, this is the first study to investigate EF in midlife adults exclusively using TMT measures. While these tasks provide valuable insights, they may not fully capture the complexity of cognitive processes affected by SNHL. It is important to note that participation in our study was limited to individuals with untreated hearing loss. Further investigation is necessary to delve into the effects of aided hearing on EF, as well as to determine whether aided hearing can ameliorate these changes. The current study utilized a cross-sectional design, limiting the ability to establish causal relationships between mild SNHL and EF deficits using the TMT. Conducting longitudinal studies to track cognitive changes over time can provide insights into the progression of EF deficits and how they relate to the severity and duration of SNHL. Such studies could also explore whether interventions such as hearing aids or cognitive training programs influence the trajectory of cognitive changes. Studies investigating cognitive function across various levels of hearing loss, ranging from mild to severe can better understand how auditory impairment impacts executive functioning. The study may have also had specific sample characteristics that limit the generalizability of findings to broader populations. A comprehensive understanding of the alterations found in the present study necessitates a multidisciplinary investigative approach encompassing behavioral, neuroimaging, and electrophysiological methodologies.

Conclusion

This study delves into the cognitive consequences of SNHL in mid-life adults, focusing on executive functions. While the research confirms that SNHL is linked to overall cognitive disturbances, the impact is not exclusive to executive function deficits, suggesting a broader cognitive influence. The results obtained in the present study accentuate the importance of early intervention for mid-life hearing loss, shedding light on a potential avenue for preventing cognitive decline and mitigating the future burden of dementia.