Calorie Restriction and Brain Function in Mild Cognitive Impairment (5-2CR)

Problems with blood sugar metabolism (i.e., metabolic dysfunction) progressively develop through old age, which is primarily due to obesity and lack of physical activity. Metabolic dysfunction increases the risk for Alzheimer's disease (AD) and negatively impacts memory and related brain function. There is intense interest in developing interventions, particularly non-drug therapies, to combat AD. Recent clinical trials have found that intranasal insulin, which facilitates glucose metabolism in the brain, is able to maintain memory in participants with Mild Cognitive Impairment (MCI), the precursor to AD. While intranasal insulin is a useful, proof-of-concept intervention, it does not affect visceral fat mass and therefore metabolic dysfunction will persist in a given person. The investigators wish to engage participants with MCI in intermittent calorie restriction (CR), to reduce metabolic dysfunction and improve glucose metabolism. Intermittent calorie restriction in this case refers to eating whatever one wants for 5 days, followed by 2 consecutive days of consuming 530 calories via one protein shake with sufficient nutrients to sustain the person. This results in reliable weight loss, which itself improves glucose metabolism in the body and has a wealth of other benefits. (It should be mentioned here that weight maintenance has been shown in studies when participants restrict to 1 day/week).

AD diagnosis is projected to increase from approximately 5 million to 13.8 million Americans by 2050. The Alzheimer's Association estimates that healthcare costs for AD by 2050 could be 1.2 trillion dollars per year for Americans aged 65 years or older. There is marked focus on treating AD during the MCI phase, which precedes AD. In the brain, insulin normally facilitates microvascular blood flow, glucose uptake, and glucose oxidation for adenosine triphosphate (ATP) generation. Insulin resistance (IR) is defined as a reduced cellular responsiveness to insulin, characterized by higher insulin levels needed to maintain glucose regulation in the periphery and certain brain areas. IR is found in MCI and AD patient brains. AD-related neuropathology, such as amyloid beta-containing plaques, progressive atrophy, and glucose hypometabolism first occur in brain areas that also have a high density of insulin receptors. Such areas include medial temporal lobe (MTL) and prefrontal cortex (PFC). Furthermore, as illustrated in the literature and previous work of the investigators, peripheral IR is associated with AD-like changes in MTL and PFC, including: 1) brain atrophy; 2) less glucose uptake; 3) accumulation of amyloid-beta, a hallmark of AD; and 4) increased phosphorylation of tau fibrils, another hallmark of AD. Finally, higher IR is related to deficits in memory performance and executive function. These cognitive deficits can be ameliorated with 40 IU of intranasal insulin, which increases insulin processing in the brain with minimal peripheral effects, where MCI and AD patients show stable visuospatial working memory, as well as declarative learning and memory. One limitation of intranasal insulin is that it does not change obesity, which causes IR, and may therefore be only temporarily effective. As such, the investigators are interested in dietary regimens that can lower IR and may have long-term beneficial effects on AD neuropathology and cognitive output. Critically, intermittent calorie restriction (CR) diets reliably decrease body weight and IR in human adults. Intermittent CR protects neurons against dysfunction and degeneration in AD models. The underlying cellular and molecular mechanisms involve improved cellular bioenergetics and up-regulation of protein chaperones and antioxidant pathways in neurons. For example, 5-2 CR, a diet consisting of 5 days ad libitum followed by 2 consecutive days consuming 500-600 kcal, decreases IR beyond what is achieved with daily CR, and has a compliance rate of 83% versus 55% even by 6 months. It is also important to mention that no serious Adverse Events have occurred during past 5-2 CR studies.

Dietary Supplement: Meal-Replacement Therapy To examine the effect of meal-replacement therapy on weight loss, behavior, and brain outcomes in overweight to obese aged adults with mild cognitive impairment.

Participants will be asked to consume a single 530 kilocalorie shake (i.e., "High Calorie" Boost shake) on a given day for two consecutive days each week. Participants will eat ad libitum during the remaining 5 days. Participants will also receive Registered Dietitian of Nutrition (RDN) consultations about dietary modifications to induce moderate weight loss. Participants will utilize these recommendations in addition to shake consumption. Subjects are also asked to follow RDN dietary recommendations.

A Registered Dietitian of Nutrition (RDN) will consult with subjects about dietary modifications to induce moderate weight loss. Participants will utilize these recommendations.

The meal replacement shake, consumed once per day on two consecutive days, will over a given week lead to negative energy balance and induce weight loss.

Degree of network strength in neural networks assessed at rest using Functional Magnetic Resonance Imaging

The construct for this total score is global cognitive performance, based on performance over a series of cognitive sub-domains. The unit is correct performance on a series of cognitive sub-domain tasks. The range is 0-30. A higher number is better.

The construct for this composite score is overall impairment in various executive function domains. The unit is the total number of errors committed during several NIH EXAMINER tasks (Continuous Performance Task, Fluency, Flanker, Set Shifting). The range is 0-43. A higher score is worse.

The construct for this factor is Verbal Memory. There are three sub-scales of main interest with the following units: 1) the total number of correctly recalled items in a 16-item list for a given trial where that list is read, spanning 5 trials ("CVLT Trials 1-5 Total"); 2) the total number of correctly recalled items in the same 16-item list after a 5 minute delay ("CVLT Short Delay"); and 3) the total number of correctly recall items in the same 16-item list after a 20 minute delay ("CVLT Long Delay"). The ranges are 0-80 for CVLT Trials 1-5 Total, 0-16 for CVLT Short Delay, and 0-16 for CVLT Long Delay. A higher number is better.

This construct probes working memory, a sub-domain of executive function. It requires participants to remember a list of numbers that gradually has more numbers added.

The construct for this test is Phonemic Fluency. The unit is total number of words beginning with the letter 'F' produced in 60 seconds. The value range is variable but is typically 0-36. A higher score is better. There are two sub-scales: 1) Total Repetitions; and 2) Total Rule Violations. The value range is variable for either sub-scale. A higher value for either sub-scale is worse.

The construct for this test is Phonemic Fluency. The unit is total number of words produced beginning with the letter 'L' produced in 60 seconds. The value range is variable but is typically 0-30. A higher score is better.

The construct for this test is Categorical Fluency. The unit is total number of words produced related to animals in 60 seconds. The value range is variable but is typically 0-40. A higher score is better.

The construct for this test is Categorical Fluency. The unit is total number of words produced related to vegetables in 60 seconds. The value range is variable but is typically 0-30. A higher score is better.

The construct for this test is Insight. The unit, in terms of % (0-100) is how well the participant thought they performed on the Fluency tasks relative to the population. We derive a sub-score that calculates the discrepancy between the stated % and the participant's actual % along the distribution. A larger discrepancy between %'s indicates worse performance.

The construct for this test is Working Memory. The unit is the total number of correct responses over 30 trials, regarding if a square on the screen was in that position 1 trial ago. The range is 0-30. A higher score is better. A D-prime performance scale is also automatically computed, based on the difference between the z-transforms of the "hit rate" and the "false positive" rate. The range is typically -0.25 to 3.67. A higher score is better.

The construct for this test is Working Memory. The unit is the total number of correct responses over 90 trials, regarding if a square on the screen was in that position 2 trials ago. The range is 0-90. A higher score is better. A D-prime performance scale is also automatically computed, based on the difference between the z-transforms of the "hit rate" and the "false positive" rate. The range is typically -1.12 to 2.48. A higher score is better.

The construct for this test is Inhibition. The unit is the total number of correct eye movements toward or away from a visual stimulus on the screen. The range is 0-40. A higher score is better.

The construct for this test is Inhibition. The units are the mean % accuracy and reaction time for choosing whether or not a symbol on the screen is the target symbol of interest (i.e., a "star"), which occurs during 100 trials. The accuracy range is 0-100%. The reaction time range is 400-2800 milliseconds. Higher accuracy is better and higher reaction time is worse. A sub-scale, False Alarm Errors, is also created. The unit is the number of times the participant makes the wrong selection during 100 trials. The range is from 0-100. A higher score is worse.

The construct for this test is Working Memory. The unit is the total number of correct number selections, where a participant selects the number of blue circles shown over 2-7 separate displays for a given trial, with a total of six trials conducted. The range is 0-27. A higher score is better.

The construct for this test is Set-Shifting. The unit is the mean % accuracy and reaction time, over 104 trials, for selecting either the proper matching color or shape of a stimulus presented at the top of the screen when given a specific prompt (i.e., "shape" or "color"). The accuracy range is 0-100% and the reaction time range is 400-2800 milliseconds. A higher accuracy % is better and a higher reaction time is worse. A sub-scale is automatically computed that weights median reaction time and accuracy into a composite score. The range is 0-10. A higher number is better.

The constructs for this test are Planning and Organizing. The unit is the total number of points that a participant accrues by completing a series of "high point" and "low point" puzzles within 6 minutes. The range is 0-1478. A higher score is better. Several sub-scales are automatically created, where a higher number is better: 1) the number of "high point" puzzles completed with a range of 0-21; 2) the number of low value puzzles completed with a range of 0-51; 3) The number of "high point" puzzles attempted with a range of 0-21; and 4) The number of "low point" puzzles attempted with a range of 0-51. Finally, a weighted composite score is created, where the unit is the value of the equation UTpct*log10(UTTotal+1), where UTpct is % of "high point" completed puzzles and UTTotal is the total number of points earned. The range is 0-2.7. A higher value is better.

The construct for this test is Inhibition. The unit is the mean % accuracy and reaction time over 24 trials, where the participant make a selection regarding where a middle arrow is facing amongst several other arrows facing the same way ("congruent") or the opposite way ("incongruent"). The accuracy range is 0-100% and the reaction time range is 500-3000 milliseconds. Higher accuracy is better and higher reaction time is worse. A composite sub-scale of performance is created basically based on log-rescaled median reaction time and accuracy. The range is 0-10. A higher number is better. Other sub-scales include accuracy and reaction time for congruent and incongruent trials. Ranges are variable. Higher accuracy is better and higher reaction time is worse.

The construct for this test is Social Behavior/Cognition. The unit is the total number of correct responses to what would generally constitute proper or improper responses to various situations in mainstream heteronormative White culture in the United States. The range is 0-24. A higher score is better.

The construct for this scale is Social Behavior/Cognition. The unit is a total score over 9 behavioral domains (e.g., "agitation") shown by the participant during NIH EXAMINER testing, where each domain is scored on a 1-4 Likert scale of "absent" to "severe" based on operationalized criteria. The range is 0-36. A higher score is worse.

The construct for this test is Attention. The unit is elapsed time taken to complete the test. The range is 10-300 seconds. A higher score is worse.

The constructs for this test are Attention and Working Memory. The unit is elapsed time taken to complete the test. The range is 10-300 seconds. A higher score is worse.

The construct for this test is Working Memory. The unit is the total number of correct responses in remembering a 2-7 number sequence and relating the sequence either forward or backward. The range is 0-28. A higher score is better.

The construct for this test is Fluid Intelligence. The unit is the total number of correct symbol-to-number substitutions made over 93 symbol stimuli presented on a piece of paper. The range is 0-93. A higher score is better.

The construct for this test is Visual Memory. The unit is the total number of correctly drawn tableaus, each composed of arbitrary geometric shapes, that are presented to the participant for 10 seconds. The range is 0-10. A higher score is better.

The construct for this total score is global cognitive performance over several sub-domains, some of which vary from the Mini-Mental State Examination. The unit is correct performance on a series of sub-domain tasks. The range is 0-30. A higher number is better.

The construct for this Likert-scaled structured interview is global cognition and overall function, where the interview is conducted with the participant and a caretaker/informant. The unit is the sum of six Likert-scaled sub-domains of cognition and function ("CDR-sum of boxes"), where on the Likert-scale a 0 corresponds to cognitively unimpaired, 0.5 suggests Mild Cognitive Impairment, and 1-3 indicate mild to severe dementia. The range is 0-18. A higher score is worse. Six sub-domain scores (Memory, Orientation, Judgment and Problem Solving, Community Affairs, Home and Hobbies, Personal Care) are also tallied. The range is 0-3 for each sub-domain. A higher score is worse.

The construct for this factor is Fluency, based on overall performance during 4 NIH EXAMINER Fluency tasks. The unit is a mean Z-Score calculated from 4 separate Z-scores based on the total correct for 4 Fluency NIH EXAMINER tasks. The range is -3 to 3. A higher score is better.

The construct for this composite score is Inhibition, based on overall performance during the NIH EXAMINER Flanker, Set-Shifting, and Saccade/Anti-Saccade tasks, as well as the total number of dysexecutive errors committed. The unit is a mean Z-score of Z-scored performance for those tests listed and a Z-score of total errors committed. The range is -3 to 3. A higher score is better.

The construct for this factor is Working Memory, based on overall performance during NIH EXAMINER tasks including Dot Counting and d-prime scales from the spatial 1- and 2-back tasks. The unit is a mean Z-score based on Z-scored values for total correct in Dot Counting and derived d-prime in each N-back task. The range is -3 to 3. A higher number is better.

Inclusion Criteria: Years of age 70-85; Body Mass Index => 28 and < 40 and weight < 350 pounds; Mini-Mental State Examination >= 24; A subjective memory concern from the participant, caregiver/informant, or a clinician; An education-adjusted score for the Logical Memory II, paragraph A Wechsler Scale (16 years: 8-10; 8-15 years: 5-7; < 0-7 years: 2-3); A Clinical Dementia Rating - sum of boxes = 0.5; No deficits in activities of daily living; Consensus confirmation of MCI by senior investigators; An informant/caregiver that is informed about the study and will be present at study visits. Exclusion Criteria: A history of a major cardiovascular event(e.g., heart attack); A history of cerebrovascular or other neurological disorders (e.g., stroke, epilepsy, multiple sclerosis, etc.); A history of substance abuse in the past 6 months; A history of eating disorders; A history of clinically significant endocrine disorders (e.g., hypothyroidism); A history of mood and anxiety disorders (e.g., Major Depressive Disorder, Generalized Anxiety Disorder, Bipolar disorder, etc.); Taking insulin; Current use of systemic corticosteroids; Current use of Warfarin; Having smoked tobacco products in the last 3 months; Contraindications for MRI (pacemakers, ferrous metal implants or shrapnel about the face or eyes, etc.); Other medical history that, in the opinion of the investigators, would constitute risk for subject participation; Hypertension, defined by a diastolic/systolic blood pressure of > 160/110; No current use of memantine or other Alzheimer's disease drugs.