**7. Functional neuroimaging and exercise**

240 Neuroimaging – Cognitive and Clinical Neuroscience

reported to be non-significant, but considered collectively, these small biases could contribute

Sociological studies have shown that women outlive men by 4 to 10 years, thereby partially explaining why there are usually more women in research studies involving older adults. Although women tend to live about a decade longer than men, they also experience an accelerated pace of physiological decline between their seventh and eighth decade of life. Older men tend to weigh more, be more physically active and have a higher degree of aerobic fitness than older women of the same age (Spriduso et al. 2005) and women's brain volumes are smaller than men (Allen et al., 2003). Thus care must be taken when studying variables with inherent age or gender differences. Colcombe et al. (2006) made no mention of controlling for age or gender in the statistical analyses of their data in order to determine if improvements attributed to aerobic fitness change was independent of age or gender influences. It is well known that both age and gender can impact brain volumes and cognition independently (Madden et al, 2009). Failure to control for these variables can lead to potentially erroneous conclusions. For example, Marks et al. (2010) initially noted moderate positive relationships in the anterior cingulum segment between cerebral white matter integrity and aerobic fitness as measured by diffusion tensor imaging (DTI). However, upon controlling for both age and gender, only the middle and posterior cingulum segments remained significantly related to aerobic fitness. Similarly, this pattern of reduced significance was repeated in a voxel-wise brain analysis on the same data (Liu et al. 2009). Thus it is critical to control for factors that are known to impact the brain and/or aerobic fitness parameters. Lastly, neither the exercise test protocol nor the "stretching and toning" prescription was ever fully described by Colcombe et al. (2006). This lack of information makes it difficult to determine the validity of the aerobic fitness and strength training outcomes, and it is even more difficult to impart health recommendations with confidence. Hence the encouraging conclusions regarding aerobic fitness and brain improvement from this study by Colcombe et al. (2006) must be viewed with cautious optimism. There are currently a few new NIH-funded exercise trials involving both healthy and diseased older adults in progress (http://projectreporter.nih.gov/reporter.cfm), with intervention timelines and exercise protocols seemingly mirroring Colcombe et al. 's initial study (2006) . Hopefully, these newer studies will not only control for potential confounding variables but also provide sufficient

to confounding the interpretation of the results.

exercise testing and training details to render their studies replicable.

Magnetic resonance angiography (MRA) in conjunction with DTI is helpful in determining the status of one's cerebral blood vessels. Using a process known as arterial spin labeling (ASL), the quality and quantity of the cerebral blood flow can be determined with or without perfusion. It is believed that the progressive reduction in cerebral blood flow attributed to the aging process may be caused by a reduced metabolic demand due to a reduction in neurotransmitter synthesis (Orlandi & Murri, 1996) and/or underlying microvascular disease (Bullitt et al., 2010). The consequential neural atrophy results in smaller cerebral arteries, increased intracranial resistance and slower arteriole vasomotor reactivity (Orlandi & Murri, 1996). Even though studies suggested aging may be associated with smaller cerebral vessels, Bullitt et al. (2010) reported that vessel diameter reductions may be compensated for by an increase in vessel number and that both larger and smaller vessels were impacted. In a sub-study comparing active versus inactive older adults, Bullitt et al. (2009) reported significantly lower vessel tortuosity along with a higher number of

**6.4 Magnetic resonance angiography and exercise impact** 

If one is interested in determining which regions of the brain are being activated /oxygenated during an exercise or cognitve task, a functional MRI (fMRI) using the BOLD response would be needed. For exercise studies, the obvious hurdle to overcome is movement as most movement causes disruption in the scanning process and poor images are created. Whereas cognitive psychology has forged numerous research pathways using fMRI with BOLD contrasts to determine regions of activation in the brain during various cognitive tasks, this has not been the case with exercise training interventions. Clearly there is a need for this type of research if one desires to investigate changes in cerebral blood flow or neural hormonal factors due to an exercise stimulus from either an acute exercise bout or in response to a chronic adaptation. The stumbling block to overcome is the exercise test itself. Most exercise studies use upright testing protocols on equipment that are large and bulky with both metal and electronic parts. All of this precludes testing within the scanning room due to the magnetic field. Furthermore, by the time the subject could be transferred from the exercise apparatus to the scanning bed, critical time would be lost such that the exercise impact on the cerebrovascular system would likely be missed in all but the most deconditioned subjects.

#### **7.1 MRI-friendly leg cycle ergometer**

Therefore, up to this point, the more feasible methodology for cerebral blood flow investigations with exercise have been with using transcranial dopplers (TCD) and/or electroencephalography (EEG). Although these methods also have difficulty with accurate measures during movement, they are in comparison, lower in cost and easier to administer than an fMRI study. However, for approximately \$75,000 (US\$), the Lode MRI-compatible recumbent leg cycle ergometry system can now be purchased from ELECTRAMED Corporation, located in Flint, Michigan, http://www.electramed.com/MRI%20ERGOMETER %20CARDIOLOGY%20\_Details.htm. This would enable the researcher to conduct exercise tests while the subject remains in the MRI unit. Also available are MRI-compatible electrocardiography and blood pressure measurement units, thereby solving the equipment issue. Unfortunately, this particular equipment model is only compatible with a 1.5 T MRI scanner and only with select manufacturers. Given that most research is now being done on 3.0 T or 4.0 T scanners, this ergometer may not be useable for many research protocols. The final issue left to resolve is an acceptable exercise protocol that would be taxing enough yet involve minimal movement from the torso up during scanning sequences. One potential resolution would be to develop an intermittent exercise test protocol so that exercise bouts would take place during the imaging sequence changes, akin to an event-related design. Since stimuli in an event-related design are presented as isolated events of short duration, a brief

MRI Techniques to Evaluate Exercise Impact on the Aging Human Brain 243

exercise in the research design should obtain the ACSM Guidelines for Exercise Testing and Training (2009). A good textbook is Nieman's (2010) exercise prescription textbook used for training undergraduates in exercise science. However, there are times when it is inconvenient, illogical, or cost-prohibitive to conduct a fitness test. For those times when actual exercise testing is ruled out, there is a rather good non-exercise aerobic fitness estimation equation that is quite easy to use, providing one is trained in obtaining a valid physical activity recall. While it can be difficult to get accurate physical activity recalls beyond a few weeks, a seasoned investigator in physical activity recall questionnaires can

One physical activity recall formula for estimating aerobic fitness has been in the literature since 1990. It was gathering dust until recently when we used it for a retrospective analysis exploring the role aerobic fitness might have on cerebral white matter integrity on both younger and older adults (Marks et al., 2007). Ever since that publication, we have been getting inquiries about the formula and how to use it. The formula was developed and tested at the Cooper Aerobic Institute in Texas on over 2,000 U.S. Air Force personnel ranging in age from 18 to 70 years. The subject population included males and females, fit and unfit, healthy and unhealthy. The estimated aerobic fitness value (VO2) has a standard error of about 5 ml/kg/min. This formula is very good for cross-sectional, population-based studies when the purpose is to simply categorize one's fitness level. However, the error range is a bit too high and the fitness categorizing a bit too vague for pre-post research designs where VO2 change is a critical factor. For that, VO2 does need to actually be measured. Although the formula tends to underestimate the highly fit and over-estimate the very low fit, all subjects are still able to be categorized accurately into a fitness level (e.g., low fit, average fit, high fit). The estimation formula and its accompanying physical activity

**VO2 max ≈ 56.363 – (0.381 \* age) + (1.951 \* PARS) – (0.754 \* BMI) + (gender \* 10.987)**

The good news is, lines of inquiry utilizing neuroimaging are still rather novel and as such, there is much to study. The bad news is, these lines of inquiry utilizing neuroimaging are still rather novel and there is much to learn, so omissions and/or mistakes in research design are to be expected. Investigating how exercise impacts the brain is akin to the first studies investigating how exercise impacted the heart and its

A limitation in several structural imaging studies (e.g., Marks et al. 2010; Bullitt et al. 2009; Colcombe et al., 2006) was lack of cognitive function testing – it is unknown if the improved brain structures found in the more active subjects would have translated into better cognitive function. Adding cognitive testing with magnetic resonance imaging (MRI) may help detect subtle changes that the standard cognitive test batteries if used alone, cannot. If the MRI is able to detect changes, independent of the cognitive tests, a therapeutic program could be implemented at an earlier stage of decline and perhaps be more effective in

 *BMI* = body mass index = weight (kg) / (height in meters2)  *PARS* = physical activity rating scale from 0 to 7 (see Table 3)

elicit excellent responses, even recalls spanning several years.

rating scale (PARS) are contained in *Table 2* and T*able 3* below:

Table 2. Estimated Aerobic Fitness (VO2 max) (Jackson et al., 1990).

where: *Gender*: 0 = women; 1 = men

**9. Limitations and suggestions** 

related vasculature several decades ago.

reducing further impairment.

cycling set that progressed in intensity with each event presentation could be incorporated (Carter and Sheih, 2010). Obviously, much pilot work would need to be done to determine the exact power outputs required to elicit measureable BOLD signals.
