Menopause Analytical Hormonal Correlate Outcome Study (MAHCOS) and the Association to Brain Electrophysiology (P300) in a Clinical Setting

Various studies have demonstrated that increased leptin levels and obesity are inversely related to cognitive decline in menopausal women. It is hypothesized that adiposity is inversely correlated with cognitive decline, as women with increased weight are less vulnerable to diminishing cognition. However, it is increasingly observed that menopausal women, even with increased adiposity, experience significant cognitive decline. Positron emission tomography (PET) has been used to analyze cognitive function and processing in menopausal women. Evoked potentials (P300) and neurophysiologic tests have validated brain metabolism in cognitively impaired patients. Post-hoc analyses of 796 female patients entering PATH Medical Clinic, between January 4, 2009 and February 24, 2013, were performed as part of the “Menopause Analytical Hormonal Correlate Outcome Study” (MAHCOS). Patient age range was 39–76 years (46.7±0.2). P300 latency and amplitude correlated with a number of hormones: follicle stimulating hormone (FSH), luteinizing hormone (LH), estradiol, estrone, estriol, DHEA, pregnenolone, progesterone, free and total testosterone, thyroid stimulating hormone (TSH), Vitamins D 1.25 and D 25OH, leptin, and insulin-like growth factor-binding protein 3 (IGF-BP3). Corrected statistics did not reveal significant associations with P300 latency or amplitude for these hormones except for leptin plasma levels. However, factor analysis showed that FSH and LH clustered together with Vitamin D1.25 and Vitamin D25OH, P300 latency (not amplitude), and log leptin were found to be associated in the same cluster. Utilizing regression analysis, once age adjusted, leptin was the only significant predictor for latency or speed (p = 0.03) with an effect size of 0.23. Higher plasma leptin levels were associated with abnormal P300 speed (OR = 0.98). Our findings show a significant relationship of higher plasma leptin levels, potentially due to leptin resistance, and prolonged P300 latency. This suggests leptin resistance may delay electrophysiological processing of memory and attention, which appears to be the first of such an association.


Introduction
Menopausal women become deficient in multiple hormones such as estrogen, progesterone, testosterone, and dehydroepiandrosterone (DHEA) [1][2][3][4][5][6] with increases in luteinizing hormone (LH), follicle stimulating hormone (FSH), and thyroid stimulating hormone (TSH) [7][8][9][10]. All of these hormones have individual as well as inter-related functions in the human body, including pulmonary, cardiovascular, gastrointestinal, and immunological functions [11,12]. Aging is continually changing as life expectancy increases, and the decrease in cognition through normal aging processes is of primary clinical interest [6,13]. In fact, estrogen deficiency has been proposed as a cause of memory deficits in postmenopausal women [14]. There are studies that suggest that LH increases after menopause with concomitant decline in cognitive performance [15]. Chorionic gonadotropin receptors and LH occur in the brain [16]. Thus, levels of LH and FSH may increase low-density lipoprotein receptor-related protein in the brain [14,17]. Levels of FSH increase dramatically in women during and after menopause and can be lowered with estrogen therapy [14,18]. Emerging evidence suggests that high TSH levels are associated with a two-fold risk of cognitive decline as well as prevalence of anomalies in musculoskeletal systems [19][20][21].

Menopause and Leptin
Weight gain, and associated leptin level changes during menopause have been demonstrated in numerous cases, as leptin was found to have significant association with metabolic factors  when compared to resistin in pre and post-menopausal women [24][25][26][27][28][29][30][31][32][33]. In particular menopausal status is linked with increased weight gain, increased central fat mass, and abnormal lipid metabolism. One case observed a significant statistical correlation with serum leptin levels and fat mass; abdominal obesity and visceral fat area were positively correlated with serum leptin level [28]. Leptin concentration was positively correlated with BMI and weight gain in healthy menopausal women [86]. There is indeed a definitive correlation with plasma leptin levels in peri-and postmenopausal women and weight gain, which suggests that there is a dynamic interaction between plasma leptin and body composition. These changes may further indicate that a significant association with leptin levels and metabolic factors could impact other hormonal composition changes that occur during menopause. One important question that concerns menopausal women is weight gain and early cognitive decline [22]. In some studies, researchers have found that obesity-related waist circumference had a positive correlation with global cognition [23]. There have been controversial reports indicating that naturally occurring estrogen may have a beneficial effect on cognition in women [33][34][35][36][37]. There is a strong relationship between the hormone leptin and obesity especially in females [36][37][38][39]; however, leptin's role in cognition has been studied with mixed results [28][29][30] and there is a notable lack of published data on P300 utilization and its relationship with plasma leptin levels, despite studies indicating the association between leptin and cognitive genes. The P300 evoked potential measures electrical brain waves and their function in particular regions of the brain. It is a diagnostic tool used to identify neurological deficits or abnormalities, such as cognitive decline and neurotransmitter deficiencies.
Moreover delayed P300 latency and obesity are both associated with dementia. Leptin resistance may delay electrophysiological processing of memory and attention. [87].

The Clinical Significance of P300 Testing
When used in a clinical setting, the P300 serves as an invaluable resource towards early diagnosis and treatment of poor cognition and loss of brain vitality due to menopausal transition. Analysis of the P300 and standard neuropsychological variables to assess patients with mild cognitive impairment (MCI) and Alzheimer's disease can provide valuable information for the detection of MCI and AD [80]. A number of studies determine that there is justification and efficacy in employing more resources unto converting P300 as an acceptable biological marker for AD [81][82][83][84]. P300 is responsive to the deterioration of language, memory, and executive functions, thereby acting as a useful tool in evaluating cognitive function [85]. The P300 was therefore a suitable tool in evaluating the variations in cognitive function and quality in subjects.
We hypothesized that women presenting with complaints related to menopause would have a number of associated hormonal changes relating specifically to both somatic and neurological symptoms. We also hypothesized that there would be an inverse relationship between plasma leptin levels and P300 speed and voltage based on literature suggesting that as women transition into menopause, they are more likely to gain weight due to associated changes leptin levels, which have been associated with a decrease in cognitive functioning [40][41][42][43]. In this paper, we focus on the role of leptin plasma levels and brain electrophysiology as measured by P300 amplitude (voltage) and latency.

Subjects
All female patients who presented to a private clinic, at or approaching the typical age of menopause onset (age 39-59), were examined. We also extended this study to include a small number of women (n = 33) between 35 to 39 years, as well as 10 women $ 60 years. The descriptive distribution summary of each measured variable collected from January 4, 2009 to February 24, 2013 in  All subjects underwent a thorough medical evaluation including a full screen for hormones including DHEA sulfate, estradiol, estrone, FSH, LH, pregnenolone, progesterone, free and total testosterone, and TSH obtained from BioReference Laboratories, Inc. Hormone blood levels were drawn between 9:30 am and 3:00 pm. A detailed medical history was obtained including information on the stage of menopause (pre, undergoing, or post), origin of menopause, and history of hormone replacement therapy as well as anthropometric characteristics such as weight, height, BMI, body fat percentage, muscle, and bone density recorded via duelenergy x-ray absorptiometry (DXA). These anthropometric characteristics were not included in the study's final data analysis, but may be considered for future studies. A Menopause Questionnaire ( Figure 1) was given to all women (n = 796) following a preliminary screening. The quantitative section of the Menopause Questionnaire consisted of 64 questions related to symptoms of menopause. Each symptom was rated on a Likert scale of frequency from 1 (never) to 5 (always). The total number of endorsed symptoms was calculated as a gross indicator of menopausal symptomatology. Mean values of the Likert ratings also were calculated within each of 12 domains of symptoms: neurological, neuropsychiatric, neuropsychological, endocrine, pulmonary, musculoskeletal, gastrointestinal, cardiovascular, immune, dermatological, genitourinary, and gynecological. A grand mean of the Likert rating across all domains also was calculated. All reported blood levels and preliminary questionnaires were recorded within 3 weeks of each other.

Statistical Analysis
In order to examine the relationships among the variables, each and every pair of response and explanatory variables were examined using contingency analysis, bivariate scatter plots, simple linear regressions, as well as one-way analysis of variance (ANOVA) or Welch's t-test for classification variables. When the test of multiple means turned out to be statistically significant at 5% level, all pairs of means were compared using Tukey-Kramer's honestly significant difference as a post-hoc method.
Pearson product-moment correlations were calculated between hormone levels and the 12 mean domain scores, the total number of endorsed systems, and the grand mean across all 64 questions of the Menopause Questionnaire. A one-way ANOVA was performed for the origin of menopause variable for each of the 12 symptom domains, with a Bonferroni correction of p,.004 required for significance. Given the large number of domains and likely high inter-correlations between them, a factor analysis with principal components extraction and varimax rotation was performed on the 12 mean domain scores. Factor scores were generated for each patient and entered into the correlation analysis with the hormone levels. A similar factor analysis was performed on the hormone levels to reduce redundancy of highly inter-correlated values. Regression analyses were performed to predict the symptom domain score factors from the 15 hormone levels and again using the hormone level factors.

EEG, P300, and Evoked Potentials Data
Lexicor and Cognitrace were used to determine P300 potential. A total of twenty electrodes were utilized: 3 along the central sulcus, 3 in the parietal region, 2 in the parietal/temporal region, 3 in the occipital region, 2 in the frontal/temporal region, as well as 5 in the frontal region. Calibration of the two machines was completed by repeat scans. Auditory stimuli of high and low beeps are utilized by Lexicor and Cognitrace with an output of latency and amplitude, dependent on age-based preprogrammed baselines. Listed in the patients' charts and determined through computer algorithm, latency (in milliseconds) and amplitude (in microvolts) from the waveform were selected for analysis.

Preliminary Analyses
The detailed results of these preliminary analyses are not shown here due to the length, but they were used in constructing more elaborate multiple regression models and interpreting the results in the Corrected Analysis Section. In particular, the following tables summarize the statistical association of certain variables of interest.  Association between Dichotomized FSH and P300 Speed/Voltage Table 2 outlines the association between FSH and P300 speed/ voltage. Latency was correlated with FSH (p = .02), but the fit was very poor (R2 = .02). P300 speed (dichotomized latency) is shown to be uncorrelated with FSH (p = .10). Amplitude or P300 voltage was also uncorrelated with FSH (p = . 19). Table 3 outlines the association between categorized FSH range and P300 latency/amplitude (i.e., speed/voltage). P300 speed was uncorrelated with FSH over 40 mIU/mL (or under 40 mIU/mL) (p = .27). Amplitude or P300 voltage also was uncorrelated with FSH over 40 mIU/mL (or under 40 mIU/mL) (p = .18). P300 speed was uncorrelated with FSH over 80 mIU/mL (or under 80 mIU/mL) (p = .65). P300 voltage also was uncorrelated with FSH over 80 mIU/mL (or under 80 mIU/mL) (p = .20). P300 speed was uncorrelated with FSH over 100 mIU/mL (or under 100 mIU/mL) (p = .13). P300 voltage also was uncorrelated with FSH over 100 mIU/mL (or under 100 mIU/mL) (p = .32). Tables 4 and 5 outline the associations between FSH range and P300 speed/voltage for FSH under/over 80 mIU/mL. Again, P300 speed was uncorrelated with FSH for the group with FSH under 80 mIU/mL (p = .12). P300 Voltage also was uncorrelated with FSH for the group with FSH under 80 mIU/mL (p = .21). Similarly, P300 speed was uncorrelated with FSH for the group with FSH over 80 mIU/mL (p = .36). Amplitude also was uncorrelated with FSH for the group with FSH over 80 mIU/ mL (p = .10). Tables 6 and 7 outline the associations between FSH range and P300 speed/voltage for FSH under/over 100 mIU/mL. Again, P300 speed was uncorrelated with FSH for the group with FSH under 100 mIU/mL (p = .29). P300 voltage also was uncorrelated with FSH for the group with FSH under 100 mIU/mL (p = .13). Similarly, P300 speed was uncorrelated with FSH for the group with FSH over 100 mIU/mL (p = .63). Amplitude also was uncorrelated with FSH for the group with FSH over 100 mIU/ mL (p = .51).

Association between FSH and P300 Speed/Voltage for FSH under/over 100
Association between FSH over 40 and Estradiol, Estrone, and Estriol Table 8 outlines the associations between FSH over 40 mIU/ mL and estradiol, estrone and estriol. There was a strong association between estradiol and FSH over 40 mIU/mL (or under 40 mIU/mL) (p,.0001). More explicitly, estradiol level was significantly higher for the group with FSH under 40 mIU/mL (126.068.6 pg/mL) than for the group with FSH over 40 mIU/ mL (27.1613.5 pg/mL) (p,.0001). There was a higher percentage of patients with the pre-menopausal level of estradiol in the group with FSH under 40 mIU/mL than in the group with FSH over 40 mIU/mL (p,.0001).
Similarly, there was a strong association between estrone and FSH over 40 mIU/mL (or under 40 mIU/mL) (p,.0001). More explicitly, estrone level was significantly higher for the group with FSH under 40 mIU/mL (82.965.0 pg/mL) than for the group with FSH over 40 mIU/mL (41.167.5 pg/mL) (p,.0001). There was a higher percentage of patients with the pre-menopausal level of estrone in the group with FSH under 40 mIU/mL than in the group with FSH over 40 mIU/mL (p,.0001).
There was a weak association between estriol and FSH over 40 mIU/mL (or under 40 mIU/mL) (p = .03). Estriol level  appeared to be slightly higher for the group with FSH over 40 mIU/mL (0.2060.01 ng/mL) than for the group with FSH under 40 mIU/mL (0.1860.01 ng/mL). After dichotomization, however, menopausal level of estriol was found to be uncorrelated with FSH over 40 mIU/mL (or under 40 mIU/mL) (p = .12).
Tables 9 and 10 outline the association between FSH and estradiol, estrone, and estriol for FSH under and over 40 mIU/ mL. Estriol was uncorrelated with FSH for the group with FSH under 40 mIU/mL (p = .61). Estriol was uncorrelated with FSH for the group with FSH over 40 mIU/mL either (p = .92). Estradiol was strongly associated with FSH for the group with FSH over 40 mIU/mL (p,.0001). Estrone also was associated with FSH for the group with FSH over 40 mIU/mL (p,.01). Table 11 outlines the association between leptin and P300 speed/voltage. High leptin levels were determined as.5 ng/mL and high P300 latencies were bdetermined by a patient's age+ 300 ms (i.e., a patient age 50 should have a P300 latency of 350, so any value greater than 350, shows a delay in latency). P300 speed appeared to be correlated with leptin (p = .01), with higher leptin level associated with abnormal P300 speed (OR = 0.97). However, once dichotomized, this association was insignificant (p = .10). Amplitude or P300 voltage was uncorrelated with leptin (p..30). Furthermore, higher leptin levels are indicative of obesity and its risks and delayed latency shows slowed brain speed and attention.

Regression Analysis
In order to assess the pattern of each hormone level on P300 speed/voltage accounting for any (additive) effect of covariate measured in this study such as age, multiple regression models were constructed with P300 speed/voltage as response variables. Final models were then built by reducing the full models through a stepwise regression method. Only the main effects were considered in the model, and dependency among the response variables was not considered for an individual examination of each response variable. The following tables and figures summarize the results of the reduced model fit, ANOVA, and effect estimates, along with prediction profiler of main effects in the reduced model.
Latency predicted by Leptin with Age (ANOVA) and P300 Speed predicted by Leptin with Age (Logistic) Once age adjusted, leptin was found to be the only statistically significant predictor for latency or P300 amplitude (p = .03), with the effect size of 0.23. Hence, the higher the leptin level, the longer the latency. After age adjustment, higher leptin level was more associated with abnormal P300 latency (OR = 0.98). Age alone also was a statistically significant factor for latency or P300 speed (p,.01): the older the patient, the longer was the latency. However, even after age adjustment, amplitude and P300 voltage were found to be uncorrelated with leptin (p..25). Also, once age adjusted, none of FSH, LH, or TSH were found to be statistically significant predictors of the P300 measures (p..25) (see Tables 12-16 and Figure 2).

Principal Component Analysis
The principal component analysis (factor analysis) was performed to derive a small number of independent linear combinations (viz., principal components) of a set of variables that capture as much of the variability in the original variables as possible. Reducing the dimensionalities of this multivariate data, the most important components were used to visualize the structure of the data. Each principal component was calculated by taking a linear combination of an eigenvector of the correlation matrix with a variable. The corresponding eigenvalues were the variance of each component.

Principal Component Analysis in the Raw Scale
The proportion of variation explained by clustering was 0.413. The first two principal components accounted for 31.7% of the variability while the first three components accounted for 43.3% of the variability. Based on these components, clustering of the variables was done producing three clusters, each with five variables. There was no distinct or definite association found among the variables (at least the ones being examined). One of the causes was the incompleteness of the data, with many missing values, along with the retrospective aspect of the study. Although some variables tended to have a sizable R2, those strongly right skewed variables from the raw data in a continuous scale were logtransformed to re-perform the principal component analysis to determine possible additional results (see Tables 17-21; Figures 3,  4).

Principal Component Analysis in the Log Scale
The first two principal components accounted for 34.1% of the variability, while the first three components accounted for 45.7%  of the variability, slightly better than the previous result in the raw scale. Based on these components, clustering of the variables was done producing five clusters, each with a varying number of variables. With an increased number of clusters, each cluster had even fewer associated variables. This confirmed the findings of no clear associations among the variables (among those being examined) and the results of regression analyses. Even in the log scale, some variables were still clustered together. FSH and LH were found to be associated while Vitamin D 1,25 and Vitamin D 25 OH tended to be together. Also, P300 latency and log leptin were found to be associated in the same cluster (see Tables 22-25; Figures 5, 6).

K-Means Clustering Analysis
As an alternative clustering analysis, the K-means clustering analysis with K = 2 was performed in the log scale of selected variables, which were strongly right skewed. Initially, the hierarchical clustering was performed to produce a tree-like classification structure, but no distinctive branches could be located to cluster the data in an effective way. In the K-means clustering, many different numbers of clusters were tried, but the two-means seemed to produce the most reasonable outcome, given the incompleteness of the dataset. The variables were scaled individually, and the within-cluster standard deviations were used producing cubic clustering criterion = 24.9793. Each cluster contained 31 and 45 points, respectively. The most representative variable separating these two clusters was log FSH, followed by log LH. The data resolution was found to be weak over other variables; see the scatterplot matrix below. With aforementioned reasons, nevertheless, the clustering was not highly successful with this dataset as the points were scattered without distinctive associations. Although the data were forced to cluster into two here, the parallel coordinate plots showed that the variable patterns were quite similar between the two groups, comparable to the overall cluster means, which resulted in low resolution of clustering (see Tables 26-30

Discussion
Although our data did not support a significant association between sex hormones and the electrophysiological parameters examined, most importantly, we did find a significant association between plasma increased leptin levels and P300 latency. Leptin was found to be the only statistically significant predictor for latency or P300 speed. Thus, the higher the leptin level was, the longer was the latency. After age adjustment, higher leptin level was associated more with abnormal P300 speed. While we found that age alone was a statistically significant factor for latency or P300 speed, we inferred that the older the patient was, the longer was the latency. However, even after age adjustment, amplitude   and P300 voltage were found to be uncorrelated with leptin. Also, once age was adjusted, none of the FSH, LH, and TSH levels were found to be statistically significant predictors for P300 speed and voltage despite some research suggesting otherwise. In particular, Davis et al. [19] found that triiodothyronine (T 3 ) levels were much lower in advanced AD Braak stages V-VI. Short et al. [15] also found significantly increased levels of LH and GRH in estrogenfree AD patients. In spite of our results, there is a historical positive correlation between the fluctuation in specific sex hormones (e.g. T3, LH, GRH, etc.) and advanced stages of AD, which should be further investigated in future studies elucidating such effects. There is also significant evidence showing that P300 evoked potentials and neurophysiological tests validate PET measures of brain metabolism in cognitively impaired patients [73]. P300 has important clinical utility in that its measurements are representative of cognitive decline, especially those patients who are already suffering from progressive cognitive impairments or dementia. Reduced P300 amplitudes, delayed latencies, and neuropsychological deficits show positive correlations with PET brain hypometabolism; thus, the P300 serves as a cost-effective indicator that is essential in primary care settings and may help determine early abnormalities in cognition.
These findings may be useful in determining the role of leptin in future studies focusing on obesity and cognition. Certainly there is enough evidence to link leptin plasma levels as a function of cellular resistance and obesity [44], especially in menopausal women [45]. While this is fairly clear, less is known concerning the role of leptin and cognition.
There have been studies suggesting that obesity and excessive weight in middle aged adults is a risk factor for cognitive decline later in life [46][47][48]; congruently, observational studies of Alzheimer's Disease (AD) patients provides potential links between adipose tissue metabolism and cognition, weight loss occurring some years prior to the onset of clinical AD symptoms [49,50]. The adipokine leptin, a cytokine secreted by adipose tissue, serves as an appetite suppressor and helps to regulate the body's energy expenditure-increased plasma levels of leptin associated with excessive body fat [51].
Mouse models have linked disrupted leptin receptors as potential sources for long-term potentiation impairments, synaptic plasticity issues, and deficiencies in spatial learning [52]. Furthermore, hippocampal viral vector introduction of leptin in rodents induced neuronal precursor proliferation and reduced neurodegeneration; in addition, it suppressed memory consolidation and reduced the spatial awareness for food of administered rodents [53,59]. The creation and removal of beta amyloid (Ab) in rodents occurs through leptin modulation [54,55], Ab hippocampal removal was related to advanced object recognition [46]. In addition, following ischemic challenge, neuronal survival in the hippocampus for Gerbils treated with leptin was greater than for gerbils without leptin treatment [56]. Senescence-accelerated prone mice (SAMP8) with impaired learning and memory showed improved retention after hippocampal leptin injections [57]. Through such rodent modeling, much has been learnt about leptin and cognition.
In humans, there remains controversy regarding the role of leptin in cognition. Lieb et al. indicated that the presence of increased hippocampal and whole brain volume were associated with higher plasma levels of leptin, suggesting that these higher plasma leptin levels associated with reduced incidence of AD [60]. Controvertibly, dementia and AD development in older adults has been linked to low plasma leptin levels [61]. Some studies have proved inconclusive regarding a difference in leptin between AD patients and controls, unless an ApoE4 allele was present [62].
Clinical attention to leptin has indicated that there is a statistical trend for increased cognitive scores to be associated with lower plasma leptin levels [58]. Interestingly, when stratified by race and gender, excessive leptin in white men was associated with higher cognitive scores, as assessed by the Montreal Cognitive Assessment, while the opposite trend persisted amongst black men [58]. AD patients with excessive low weight and decreased leptin plasma levels, implicates a hypothalamic level malfunction [63]. Using the Modified Mini Mental State Exam, there is shown to be a protective quality of cellular leptin combating cognitive decline [64,65].
Finally, the increased gender difference regarding the incidence of AD in women can potentially be linked to the increased relative loss of adipokines in women than in men. Further gender differences in leptin occur regarding C-reactive protein (CRP)  levels, highly correlated to leptin levels in women but not in men, which potentially links increased CRP levels to the increased relative adipose load for women [66]. Increased abdominal fat in Japanese men has been associated with mild cognitive impairment, the association inconclusive for Japanese women [67].
We are cognizant of the extensive literature related to the effects of a number of hormones that we studied in this article and their respective roles of influencing leptin function: LH [68,69]; FSH [70][71][72]; estradiol [61,63,72].

Limitations
Although we started with a cohort of 796, after sub analysis, the lowest N was 194 due to missing values. Therefore, we caution any definitive interpretation of these data. Of the remaining 194 subjects, only 159 subjects had leptin values available for statistical analysis against P300 values. While we are encouraging more research in this area of important investigation, we suspect that leptin resistance may have a negative influence on global cognition in women based on the limited cohort of 159 women that was investigated.
We are cognizant that P300 latency may change as a function of pharmaceutical treatments, anxiety, and other physiological and behavioral factors. Thus, these factors must be considered seriously, and we propose caution in interpreting these data. Subsequently, in future studies we will also include Mini-Mental State Examination status, Wechsler Memory Scale tests, and other neurocognitive measures [73] along with electrophysiological measurements.
In pre-and post-menopausal women, we attempted to determine the relationship between plasma leptin levels and brain functionality as measured by P300. However, a PubMed literature search on 10-27-2013 revealed no articles on P300 measures and plasma leptin levels, in spite of many references associating leptin levels and genes with cognition [74][75][76][77][78][79].
Specific demographics and an investigation of correlated BMI levels with leptin should be further expanded in future studies. We were unable to include additional demographic data such as body weight, height, and BMI among other fields. We are aware that these additional parameters may add to this study and better evaluate the changes in body composition among menopausal women in addition to cognitive function. A specialized study of this nature would be a significant follow-up to our current observations.

Conclusions
We are proposing that since leptin has been considered to be a cognitive enhancer, whereby cellular leptin acts on hippocampal cells (as one example) and improves performance in object recognition; leptin resistance and subsequent higher plasma levels should induce cognitive decline. As we expected, higher plasma leptin levels significantly and positively associated with a prolonged latency, which is a determinant for cognitive function.
Thus, our findings show significant association of higher plasma leptin levels, potentially due to leptin resistance, and prolonged P300 latency appears to be the first of such an association. Leptin resistance may delay electrophysiological processing of memory and attention.