2007 ; DeCaro 2008 ; Meyer and Novak 2012 ). Less direct measures use the
physiological effects of the two stress hormone systems as indicators of stress:
elevated heart rate, blood pressure, sweating on the palms, and changes in the
immune system present in small amounts of blood collected fromfinger sticks
(James and Brown 1997 ; McDade 2007 ; Bali and Jaggi 2015 ).
Some biological markers of stress are only useful in controlled laboratory or
clinical settings. As one example, functional MRIs are now distinguishing cerebral
bloodflow changes under psychological stress (Wang et al. 2005 ). Studies using the
galvanic skin response employ skin conductance to measure palmar sweating; this
measure is less useful under ambulatory conditions where physical activity and
environmental heat affects the sweat rate independent of stress. In fact, virtually all
measures that are used commonly during normal settings have some drawbacks,
requiring careful consideration of factors other than stress that may contribute to the
biological markers. For instance, changes in heart rate are an indication of
stress-induced sympathetic nervous system activation, but also are very closely
related to physical activity levels (Edholm 1967 ). Ambulatory blood pressure
changes are also both an indicator of stress and of physical activity. When ambu-
latory blood pressure is used as a stress measure, studies need to account for
physical activity; this can be accomplished through concurrent self-reports of
activity in diaries, monitoring of heart rate (heart rate is more closely connected to
physical activity levels than is blood pressure), or use of accelerometry (Van Egeren
1991 ; Gretler et al. 1993 ; Leary et al. 2000 ). Even measurements using nore-
pinephrine and cortisol require consideration of physical activity, since activity
affects these hormone levels (Pearson et al. 1995 ).
The Problem of Poor Correlations Among Stress Markers
There are strong correlations between some of the biological markers of stress, but
these correlations vary depending on circumstances. Levels of salivary
alpha-amylase are significantly correlated with levels of plasma epinephrine and
norepinephrine in some situations but not in others (Chatterton et al. 1996 ;
Rohleder et al. 2004 ; Nater et al. 2006 ). Physiologically, there is a strong inter-
connection between the sympathetic adrenal medullary and pituitary-adrenal cor-
tical systems (Brown 2007 ). While studies have shown correlations between urinary
catecholamines and cortisol when people are exposed to stress (Morgan et al. 2001 ),
there are also differences (Dimsdale and Ziegler 1991 ). As noted, catecholamine
levels become elevated much quicker than do cortisol levels in response to stress.
Also, prolonged stress can lead to reduced cortisol levels, while catecholamine
levels remain high (e.g., Mason et al. 1988 ; Miller et al. 2002 ), and there may be
seasonal differences in the relative amounts of these hormones as well (Hansen
et al. 2001 ).
The variation in correlations among stress measures becomes more marked when
one compares self-report measures with biological markers of stress. In a study of
7 Stress Biomarkers as an Objective Window on Experience 123