brain research 1558 (2014) 44–56

Available online at www.sciencedirect.com

www.elsevier.com/locate/brainres

Research Report

Glucose-induced inhibition of the appetitive brain response to visual food cues in polycystic ovary syndrome patients Dean A. Van Vugta,b,n, Alicja Krzemiena, Hanin Alsaadib, Tamar C. Frankb, Robert L. Reida a

Department of Obstetrics & Gynaecology, Queen's University, Kingston, Canada Department of Biomedical and Molecular Sciences, Queen's University, Kingston, Canada

b

ar t ic l e in f o

abs tra ct

Article history:

We postulate that insulin regulation of food intake is compromised when insulin

Accepted 19 February 2014

resistance is present. In order to investigate the effect of insulin sensitivity on appetitive

Available online 26 February 2014

brain responses, we conducted functional magnetic resonance imaging studies in a group

Keywords:

of women diagnosed with polycystic ovary syndrome (PCOS) in which insulin sensitivity

Insulin resistance

ranged from normal to resistant. Subjects (n ¼19) were imaged while viewing pictures of

Glucose challenge

high calorie (HC) foods and low calorie (LC) foods after ingesting either 75 g glucose or an

Functional magnetic resonance

equivalent volume of water. The insulin sensitive group showed reduced blood oxygen

imaging

level dependent (BOLD) signal in response to food pictures following glucose ingestion in

Polycystic ovary syndrome

numerous corticolimbic brain regions, whereas the insulin resistant group did not. There was a significant interaction between insulin sensitivity (sensitive vs resistant) and condition (water vs glucose). The largest clusters identified included the left insula, bilateral limbic/parahippocampal gyrus/culmen/midbrain, bilateral limbic lobe/precuneus, and left superior/mid temporal gyrus/parietal for HC and LC stimuli combined, the left parahippocampal gyrus/fusiform/pulvinar/midbrain for HC pictures, and the left superior/ mid temporal gyrus/parietal and middle/inferior frontal gyrus/orbitofrontal cortex for LC pictures. Furthermore, BOLD signal in the anterior cingulate, medial frontal gyrus, posterior cingulate/precuneus, and parietal cortex during a glucose challenge correlated negatively with insulin sensitivity. We conclude the PCOS women with insulin resistance have an impaired brain response to a glucose challenge. The inability of postprandial hyperinsulinemia to inhibit brain responsiveness to food cues in insulin resistant subjects may lead to greater non-homeostatic eating. & 2014 Elsevier B.V. All rights reserved.

n

Correspondence to: 3002 Etherington Hall, Queen's University, 94 Stuart St., Kingston, Ontario, Canada K7L 3N6. Fax: þ1 613 533 6779. E-mail address: [email protected] (D.A. Van Vugt).

http://dx.doi.org/10.1016/j.brainres.2014.02.037 0006-8993 & 2014 Elsevier B.V. All rights reserved.

45

brain research 1558 (2014) 44–56

1.

Introduction

Insulin is an important signal of energy balance (Benoit et al., 2004). Positive and negative energy states are associated with increased and decreased insulin concentrations respectively. Insulin concentrations in cerebrospinal fluid fluctuate in parallel with peripheral insulin concentrations (Strubbe et al., 1988). Insulin gains access to the brain through a saturable transport process via the endothelial cells of the blood brain barrier (Baura et al., 1993; Banks et al., 1997). Insulin receptors and insulin receptor substrate-2 are present in brain regions known to control food intake including the ventral tegmental area (VTA)/substantia nigra, olfactory bulb, hippocampus, hypothalamus, area postrema, medial nucleus of the solitary tract, and dorsal motor nucleus of the vagus nerve (Schulingkamp et al., 2000). Intracerebral ventricular administration of insulin reduced food intake and induced weight loss in baboons and sheep (Woods et al., 1979; Foster et al., 1991). Conversely, targeted disruption of neuronal insulin signaling resulted in hyperphagia and weight gain (Bruning et al., 2000; Obici et al., 2002). Insulin resistance results in a reduced ability of insulin to stimulate glucose uptake in muscle and fat and to inhibit gluconeogenesis in the liver. Because glucose uptake by the brain is independent of insulin action, the idea that insulin resistance extends to the brain has been dismissed. However, it has been postulated that insulin signaling in the brain may be impaired when insulin resistance is present (Plum et al., 2005; Pagotto, 2009). Insulin-induced glucose metabolism as measured by positron emission tomography was reduced in the ventral striatum and prefrontal cortex of men with insulin resistance (Anthony et al., 2006). Insulin-induced cortical activity measured by magnetoencephalography during a euglycemic clamp correlated with insulin sensitivity (Tschritter et al., 2006). Hypothalamic activity as measured by functional magnetic resonance imaging (fMRI) was inhibited by a glucose challenge in healthy men, but not in men with

type 2 diabetes (Vidarsdottir et al., 2007). Food pictureinduced activation of the anterior cingulate, dorsolateral prefrontal cortex (DLPFC), midbrain, and lateral orbitofrontal cortex in women with polycystic ovary syndrome (PCOS) was negatively correlated with insulin sensitivity (Van Vugt et al., 2013). The objective of the current study was to determine if appetitive brain responses to a glucose challenge are affected by insulin sensitivity. We postulated that a glucose challenge would reduce appetitive brain responses, but this response would be compromised in insulin resistant subjects. To accomplish this goal, we characterized the BOLD response in a group of women diagnosed with PCOS in which insulin sensitivity ranged from normal to resistant. We chose to study this question in the clinical setting of PCOS because of the high incidence of insulin resistance in PCOS (Dunaif et al., 1989; Dunaif, 1997; Lo et al., 2006).

2.

Results

2.1.

Demographics

Mean age and endocrine/metabolic measurements (7SD and range) for all subjects combined and for subjects divided into insulin sensitive and insulin resistant groups are shown in Table 1. Eight of 19 subjects had a 2 h G:Io1.5 and a HOMA2 sensitivity o60% and were designated insulin resistant. The 11 remaining subjects were classified as insulin sensitive. Eight of eleven had a 2 h G:I41.5 and a HOMA2 460%, whereas the remaining three had a 2 h G:I41.5 but a HOMA2 o60%. Based on body mass index (BMI), 6 subjects (5 sensitive) were normal or overweight (23.5–29.7 kg/m2), 7 subjects (4 sensitive) were obese (30.7–39.7 kg/m2), and 6 subjects (2 sensitive) were morbidly obese (42.6–55.5 kg/m2). Two subjects in the resistant group were prescribed metformin and one subject in the sensitive group was prescribed an oral contraceptive pill. BMI, waist circumference, 2 h G:I, HOMA2,

Table 1 – Clinical and biochemical features of study subjects.

Age BMI (kg/m2) Waist circumference (inches) 2 h OGTT G:I (mg/dl/μU/ml) HOMA2-IR (% Sensitive) Fasting glucose (mg/dl) Fasting insulin (μU/ml) Testosterone (nmol/L)

n

po0.05 resistant vs sensitive; po0.01 resistant vs sensitive; nnn po0.001 resistant vs sensitive. nn

Overall (n ¼19)

Sensitive (n ¼ 11)

Resistant (n ¼8)

27.675.24 (18.0–39.0) 36.179.16 (23.5–55.5) 41.677.8 (32.0–60.0) 2.5772.202 (0.5–8.9) 81752.8 (29–212) 89.879.72 (70.3–106.3) 12.976.51 (3.5–26.6) 2.270.82 (1.0–4.0)

26.574.72 (20.0–36.0) 32.577.75 (23.5–47.6) 37.575.32 (32.0–48.0) 3.772.35 (1.5–7.4) 109754.7 (50–212) 89.0711.65 (70.3–106.3) 8.273.87 (3.5–15.0) 2.170.91 (1.0–3.6)

29.375.80 (18.0–39.0) 41.278.93n (26.7–55.5) 47.177.40nn (35.0–60.0) 1.170.26nn (0.5–1.3) 4377.1nnn (29–53) 90.877.27 (77.5–97.3) 18.873.61nnn (14.7–26.6) 2.570.94 (1.2–4.0)

46

brain research 1558 (2014) 44–56

Table 2 – Effect of food picture exposure on hunger.

Water condition Pre-scan Post-scan

Glucose condition Pre-scan Post-scan

†

Overall (n ¼19)

Sensitive (n ¼ 11)

Resistant (n ¼ 8)

6.572.27 (2–10) 8.171.63† (5–10)

6.872.27 (5–10) 8.171.97 (5–10)

6.172.36 (2–9) 8.171.13 (6–9)

6.772.31 (1–10) 7.471.68 (4–10)

6.472.66 (1–10) 7.671.86 (4–10)

7.371.75 (5–9) 7.171.46 (5–9)

Po0.05 vs pre-scan vs post-scan; (Friedman test).

and fasting insulin were significantly different between the resistant and sensitive groups, whereas age, fasting glucose, and testosterone did not differ between the two groups.

was observed in numerous corticolimbic regions of resistant subjects to the LC–C contrast.

2.3.2. 2.2.

Behavioral results

No qualitative difference in self-reported breakfast consumption was noted between groups or conditions. The mean postscan hunger score (7SD) after water consumption was significantly increased compared to the pre-scan hunger score for groups combined (8.171.63 vs 6.572.27) (Table 2). In contrast, the post-scan hunger score after glucose consumption was not different from the pre-scan hunger score for the groups combined (7.471.68 vs 6.772.31). The difference between pre- and post- scan hunger score within the sensitive or resistant groups was not significant for either condition. Lastly, hunger scores compared between groups (pre vs pre and post vs post) showed no significant differences for either condition. The other characteristics assessed by the pre- and post-scan questionnaire were examined, but not quantified, as they were included primarily to identify anyone who might have been uncomfortable or unwell at the time of the scan and to act as a distracter from the main outcome of interest, namely hunger. Average picture recall was 74.5% correct for the groups and conditions combined. Recall was similar for the three types of pictures and was unaffected by either insulin sensitivity or condition (range was74.0–75.6%).

2.3.

Interaction

There was a significant interaction between condition and sensitivity which identifies group differences for the differences between conditions i.e. Sensitive(water–glucose)– Resistant(water–glucose) and its converse [Resistant(glucose–water)– Sensitive(glucose–water)]. A positive interaction was seen for three clusters (frontal lobe/insula, parahippocampal gyrus/culmen/ midbrain, and precuneus) for the Food (HCþLC)–C contrast (Fig. 1) and a large cluster constituted by the parahippocampal gyrus/lingual/thalamus/precuneus/fusiform/pulvinar/midbrain in response to the HC–C contrast. An interaction for the LC–C contrast was seen in three large clusters consisting of the culmen/lingual/anterior cerebellum, middle frontal gyrus/ inferior frontal gyrus/inferior orbitofrontal cortex, and superior temporal gyrus/parietal/mid temporal (see Table 4 for a complete list of clusters exhibiting an interaction between condition and sensitivity for the various contrasts). Thus, the resistant group had greater reactivity to food pictures following glucose ingestion (with activity following water ingestion subtracted) compared to the sensitive group in numerous regions. The interaction between condition and insulin sensitivity on brain responsiveness to food cues is further illustrated by plotting the parameter estimates (Fig. 2). Mean signal intensity to food pictures was reduced by glucose compared to water in the sensitive group, but increased by glucose compared to water in the resistant group.

Imaging results 2.3.3.

2.3.1. Comparison of brain reactivity in insulin sensitive and insulin resistant subjects A differential response between water and glucose to the various contrasts differed in insulin sensitive and insulin resistant subjects (Table 3). A differential BOLD response (water4glucose) to the HC–C and HC–LC contrasts was observed in numerous corticolimbic brain regions of insulin sensitive subjects, whereas no regions exhibited a greater response to glucose compared to water. A very different response was observed in the insulin resistant group. HC–C and HC–LC contrasts failed to elicit greater BOLD responses during water compared to glucose in the resistant group. Furthermore, the opposite differential response (glucose4water)

Correlation with insulin sensitivity

An effect of insulin sensitivity on brain responsiveness to visual food cues during a glucose challenge was examined by including 2 h G:I as a regressor. This analysis identified a number of clusters whose BOLD signal was negatively correlated with insulin sensitivity. Most notably, activity in the posterior cingulate/precuneus and in a large cluster constituted by the medial frontal gyrus/anterior cingulate in response to the HC–C contrast was negatively correlated with insulin sensitivity (Fig. 3). Table 5 contains a complete list to this and the other 2 contrasts. The only positive correlation seen was for the lentiform/putamen/pallidum to the HC–LC contrast. Inclusion of BMI as a nuisance covariate did not detract from the relationship between insulin sensitivity and

47

brain research 1558 (2014) 44–56

Table 3 – Differential response (water vs glucose) in insulin sensitive and resistant subjects. Region

Sensitive HC–C (water4glucose) DLPFC, cingulate gyrus, limbic lobe Precuneus, posterior cingulate, limbic lobe ParaHippocampal, fusiform, lingual, culmen Vermis Hippocampus, pulvinar, parahippocampal gyrus, insula, limbic lobe Midbrain Frontal lobe Limbic lobe, parahippocampal gyrus, thalamus, lingual, brainstem, pulvinar, Midbrain Supp motor area, cingulate gyrus Cerebellum DLPFC Precuneus, posterior cingulate Thalamus, pulvinar Limbic lobe Cerebellum Middle temporal gyrus Caudate

Hem

k

tmax

MNI coordinates x

y

z

L L L L L L R R

209 126 124 155 231 82 73 223

5.6 5.5 5.4 5.3 5.2 5.1 4.8 4.8

18 6 20 4 18 6 30 8

18 52 44 70 38 20 6 32

38 20  12  16 6  16  14 4

R L R R R L R R R

50 40 159 164 57 31 21 44 25

4.8 4.5 4.4 4.3 4.3 4.2 4.1 4.1 4

14 18 32 16 26 14 46 60 14

8 60 6 60 30 14 44 18 6

48  32 44 10 14 36  36  16 10

LC–C (water4glucose) Cingulate gyrus

R

31

4.46

10

12

30

HC–LC (water4glucose) Frontal lobe Pulvinar, thalamus Insula Cerebellum, culmen, declive, tuber Cerebellum, culmen Mid cingulum Midbrain DLPFC

R R L L R L L R

69 181 161 118 186 30 39 36

6.7 5.1 5.1 5 5 4.6 4.2 4.2

24 10 28 42 2 6 6 28

22 30 26 62 54 44 24 12

52 6 28  36  10 52  22 42

Resistant HC–C (glucose4water) Cerebrum, posterior cingulate

L

80

4.66

16

38

12

L L L R R L R L R R L R L R

215 88 116 72 24 65 28 46 33 75 27 23 25 58 47 21 27

5.42 5.14 5.02 4.97 4.81 4.79 4.68 4.68 4.50 4.39 4.39 4.40 4.30 4.30 4.20 4.11 3.95

32 26 6 16 46 28 12 16 40 12 60 4 60 44 0 44 0

58 38 20 30 4 10 38 28 4 38 22 10 0 72 42 10 4

16 6 10  16 20 18  16  14  22 8 0 0  12 40  16 2 24

69

5.05

34

52

24

LC–C (glucose4water) Middle temporal gyrus Middle frontal gyrus, OFC, Inferior frontal gyrus Corpus callosum/caudate Cerebellum, parahippocampal Frontal lobe, rolandic operculum Frontal lobe/extra nuclear Medial frontal gyrus Parahippocampal Temporal lobe Posterior cingulate, hippocampus Mid temporal, sup temporal Caudate Mid temporal Parietal lobe, inferior parietal Medial frontal gyrus Insula Corpus callosum anterior cingulate HC–LC (water4glucose) Temporal lobe

L

L

Abbreviations: Hem ¼ Hemisphere, k ¼ number of voxels within a cluster, HC ¼high calorie, LC¼ low calorie, C ¼ control, DLPFC ¼dorsolateral prefrontal cortex, OFC ¼ orbitofrontal cortex.

48

brain research 1558 (2014) 44–56

3. Ins

4

AC

2

0

Z=26

STG

Thal

4 Prec 2

0

Z=16

OFC 4

2

0

PG

Z=-6

Fig. 1 – Interaction between insulin sensitivity and condition. Axial images are shown for three specified Z coordinates. A threshold of po0.005 uncorrected and 20 contiguous voxels was applied for display purposes. The color scale indicates the t value of functional activity. Activation in response to Food – C following glucose ingestion minus water ingestion was increased in the insulin resistant group compared to the insulin sensitive group in the anterior cingulate (AC) and insula (Ins; top panel), superior temporal gyrus (STG), precuneus (Prec), thalamus (Thal) and insula (middle panel), and left inferior orbitofrontal cortex (OFC) and parahippocampal gyrus (PG; bottom panel).

BOLD. Furthermore, a regression analysis between BOLD signal and BMI, unlike for 2 h G:I, identified a limited number of regions. The supplemental motor area was positively correlated with BMI for the HC–C contrast, the fusiform was negatively correlated with BMI for the LC–C contrast, and the fusiform and frontal inferior tri were positively correlated with BMI for the HC–LC contrast (data not shown).

Discussion

To our knowledge, this is the first study to determine the effects of insulin sensitivity on brain reactivity to visual food cues during a glucose challenge. We show that food pictures elicited very different responses in insulin sensitive PCOS and insulin resistant PCOS subjects during a glucose challenge compared to water. Reactivity to HC food pictures following a glucose challenge was significantly reduced in numerous regions in the insulin sensitive group, but not in the resistant group. This dichotomous response between insulin sensitive and insulin resistant subjects is further evidenced by the interaction analysis. Notable regions exhibiting an interaction include the insula, anterior cingulate, precuneus/posterior cingulate, parahippocampus/midbrain, orbitofrontal cortex, and temporal gyrus. These regions are activated by food cues, often in an energy dependent manner (fasted state 4 fed state) (LaBar et al., 2001; Killgore et al., 2003; Hinton et al., 2004; St-Onge et al., 2005; Holsen et al., 2005; Porubska et al., 2006; van der Laan et al., 2011). Differences in reactivity to food pictures following glucose ingestion may identify neural substrates that are inhibited by caloric intake in insulin sensitive subjects, but are dysfunctional when insulin resistance is present. Because these regions process food motivation and reward (Van Vugt, 2010), it is conceivable that differences in brain reactivity to food cues may affect food ingestion. Our observation of a reduced BOLD response to HC food pictures during a glucose challenge in the insulin sensitive group is in agreement with a recent study that reported an inhibitory effect of a glucose challenge on brain reactivity to palatable food pictures in normal weight subjects (Kroemer et al., 2013). BOLD signal in the fusiform gyrus, superior temporal gyrus, medial frontal gyrus, and limbic lobe in response to food pictures was reduced during an OGTT. Furthermore, the magnitude of the BOLD response correlated with the OGTT-induced change in insulin concentration (Kroemer et al., 2013). We confirmed reduced reactivity in the middle temporal gyrus, caudate, and medial frontal gyrus, but also observed reduced reactivity to HC food pictures in several other regions including the midbrain, pulvinar/insula/parahippocampal gyrus, and precuneus/posterior cingulate. Unlike Kroemer et al. (2013), we did not observe increased reactivity to HC food pictures in any brain regions during a glucose challenge in insulin sensitive subjects. These differences may relate to differences in subject characteristics (normal weight vs. overweight; male and female vs. female only) and length of food deprivation (overnight fast vs. 6 h). Kroemer et al. also reported a negative correlation between insulin concentrations induced by an OGTT and cerebellar, insular, striatal, cingular, inferior frontal, prefrontal, and temporal activation by palatable food pictures (Kroemer et al., 2013). This association was interpreted as evidence of insulin feedback inhibition. Our demonstration of a negative correlation between insulin sensitivity and the BOLD response to HC food pictures in numerous brain regions (inferior parietal, posterior cingulate/precuneus, middle frontal gyrus/anterior cingulate, precentral, temporal mid) extends this concept of insulin feedback to include the

49

brain research 1558 (2014) 44–56

Table 4 – Interaction between insulin sensitivity and condition. Region

Positive interaction Food-C Posterior cingulate, precuneus, limbic lobe Frontal Lobe, insula Limbic lobe, parahippocampal gyrus, culmen, midbrain Precuneus, limbic lobe, parahippocampal Limbic lobe, parahippocampal gyrus Frontal lobe Extra nuclear, anterior cingulate Frontal lobe, OFC Parietal lobe, precuneus, mid cingulum Mid cingulum Frontal lobe, middle frontal gyrus Limbic lobe, calcarine, precuneus Frontal lobe, insula Limbic lobe, cingulate gyrus Superior temporal gyrus, mid temporal, parietal Middle temporal gyrus

Hem

k

tmax

MNI coordinates x

y

z

L L R L L R L L L R L R R L L L

93 159 247 175 47 53 96 31 85 92 26 104 26 26 102 22

5.42 5.22 4.96 4.92 4.63 4.60 4.50 4.48 4.43 4.43 4.43 4.31 4.30 4.28 4.23 4.03

6  26 14  24  20 30 4  34  10 6  24 18 40  12  36  38

52 2 36 50 40 4 2 36 44 4 38 60 2 8 48 66

20 26 6 6  14 42 24 8 52 36 6 12 18 48 16 16

L

652

5.61

 26

46

0

R R L R L R R L L R L R R

23 52 45 47 45 47 54 37 43 29 33 31 23

5.41 4.80 4.73 4.61 4.59 4.58 4.38 4.35 4.34 4.34 4.20 4.09 4.01

34 2  10 26  38 6 30 4 4 6  12 30 20

46 68 44 42 70 4 4 54 2 36 8 48 32

2  16 52  10 14 36 42 18 24 54 50 20 10

LC–C Superior temporal gyrus, parietal, mid temporal Middle frontal gyrus, inferior frontal gyrus, OFC Culmen, lingual, cerebellum Frontal lobe, insula Posterior cingulate, precuneus Frontal lobe Anterior cingulate Mid cingulum Posterior cingulate, precuneus

L L R R L L L R R

158 102 80 35 43 56 31 27 46

5.37 5.15 5.11 5.01 4.80 4.46 4.07 3.85 3.82

 32  26 16 38 6  26 6 4 8

50 38 32 0 52 6 0 6 54

24 6  14 20 20 30 18 34 16

HC–LC Middle temporal gyrus

L

50

4.93

 38

72

14

Negative interaction HC–LC Sub gyral

L

20

4.45

 34

52

24

HC–C Parahippocampal gyrus, lingual, thalamus, precuneus, fusiform, pulvinar, midbrain Temporal lobe Cerebellum Mid cingulum, parietal, precuneus Parahippocampal gyrus, fusiform Middle temporal gyrus, Mid occipital Mid cingulum Frontal lobe Posterior cingulate, precuneus Corpus callosum Paracentral lobule Cingulate gyrus Right cerebrum Thalamus, pulvinar

Abbreviations: Hem ¼Hemisphere, k ¼ number of voxels within a cluster, HC ¼ high calorie, LC¼ low calorie, C ¼ control, OFC ¼orbitofrontal cortex.

50

brain research 1558 (2014) 44–56

Precuneus

2.5

Midbrain

2.5

2

2

1.5

a,b

a,b

1

1.5 1

0.5

0.5

0

0

-0.5

-0.5

-1

-1

-1.5

-1.5

a,b Anterior cingulate

Insula

2

a,b

2

a,b

1.5

1.5 a,b

1 0.5

1

a,b

0.5

0

0

-0.5

-0.5

-1

-1 Sensitive

Resistant

Sensitive

W/G

W/G

W/G

Resistant W/G

Fig. 2 – Effect of insulin sensitivity on brain reactivity to food pictures. BOLD signal induced by food pictures (Food – C) following water (W;’ ) or glucose (G; ) are compared in insulin sensitive (n ¼ 11) and insulin resistant subjects (n ¼ 8). Mean parameter estimate (7SEM) was calculated for a 5 mm radius sphere centered on the peak voxel in the precuneus ( 24 50 6), midbrain (14 36  6), insula (  26 2 26), and anterior cingulate (  4 2 24). Significant differences within and between groups are indicated by a and b respectively (po0.05; ANOVA and Tukey).

intriguing possibility that insulin negative feedback on appetitive brain circuits is compromised when peripheral insulin resistance is present. Previous studies using different imaging modalities have examined the effect of insulin sensitivity on brain activity. Insulin-induced glucose metabolism measured by positron emission tomography was reduced in the ventral striatum and prefrontal cortex of insulin resistant men compared to men with normal insulin sensitivity (Anthony et al., 2006). Cortical activity as measured by magnetoencephalography during a hyper-insulinemic/euglycemic clamp was increased in lean, but not obese men, and the response correlated positively with insulin sensitivity (Tschritter et al., 2006).

Hypothalamic BOLD signal was inhibited by a glucose challenge in healthy men, but not in men with type 2 diabetes (Vidarsdottir et al., 2007). Since none of these studies used food imagery, they are not directly comparable to our own. However, collectively, these studies and our own support the premise that insulin resistance affects how the brain responds to acute changes in metabolic cues such as insulin and glucose.

3.1.

Effect of adiposity

Two recent studies compared brain responses to visual food cues in lean and obese subjects while manipulating insulin or

brain research 1558 (2014) 44–56

4 2

Prec

AC

0 PC

-2 -4 -6

Z=8

4 Ling

2

MFG

0

AC

-2 -4 -6

Z=-4

4 2 0 -2

mOFC

-4 -6

Z=-12

Fig. 3 – Negative correlation between insulin sensitivity and brain activation during a glucose challenge. Axial images are shown for three specified Z coordinates. A threshold of po0.005 uncorrected and 20 contiguous voxels was applied for display purposes. The color scale indicates the t value of functional activity that correlated negatively with insulin sensitivity as defined by the 2 h glucose to insulin ratio. Identified regions included the precuneus (Prec), posterior cingulate (PC; top panel), anterior cingulate (AC), medial frontal gyrus (MFG), and lingual gyrus (Ling; middle panel), and medial orbitofrontal cortex (mOFC; bottom panel).

glucose concentrations (Page et al., 2011; Guthoff et al., 2011). One study showed that intra-nasal insulin administration increased cerebral processing of food pictures as measured by magnetoencephalography in lean, but not in obese subjects (Guthoff et al., 2011). The other study compared BOLD responses to food pictures during a hyperinsulinemic/euglycemic clamp vs a hyperinsulinemic/

51

hypoglycemic clamp. Food pictures activated the prefrontal cortex/anterior cingulate during euglycemia relative to hypoglycemia in lean, but not obese subjects, whereas the substantia nigra/ventral tegmental area, striatum, hypothalamus, and thalamus were activated during hypoglycemia relative to euglycemia in obese subjects relative to lean subjects (Page et al., 2011). Unfortunately, neither study included a measure of insulin sensitivity. These two studies differ from each other in that intra-nasal insulin increased insulin levels in the brain without affecting glucose, whereas insulin levels were kept constant (albeit elevated) while manipulating glucose in the study by Page et al. (Page et al., 2011). Therefore, differences in activation by food pictures in lean and obese subjects were ascribed to insulin in one study, but to glucose in the other. Since both insulin and glucose levels are increased by a glucose challenge, our results may be due to a change in insulin, glucose, or a combination of the two. In fact, hyperglycemia may be required for insulin signaling in the brain to affect appetite in females. Unlike males, females did not exhibit insulininduced anorexia when fasted, but did in the postprandial state (Clegg et al., 2003; Benedict et al., 2008; Hallschmid et al., 2012). Therefore, we cannot exclude the possibility that changes in glucose contributed to the differences in BOLD signal observed between insulin sensitive and resistant groups. Additionally, the possibility that ghrelin contributed to the BOLD response deserves consideration. Ghrelin is an orexigenic peptide released from the stomach. Its secretion is inhibited by food intake as well as by an OGTT, and the magnitude of the response is modulated by insulin sensitivity (le Roux et al., 2005; Paik et al., 2006; StPierre et al., 2007). Furthermore, exogenous Ghrelin has been shown to increase brain reactivity to food images (Malik et al., 2008). Several imaging studies reported differences in brain reactivity to visual food cues in obese subjects compared to normal weight subjects before and after eating (Del Parigi et al., 2002; Holsen et al., 2006; Martin et al., 2009; Dimitropoulos et al., 2012) or after a glucose challenge (Miller et al., 2007). Whereas healthy weight controls showed greater activation to food pictures in the amygdala, orbitofrontal cortex, medial prefrontal cortex, and frontal operculum before eating compared to after eating, Prader–Willi Syndrome patients had greater activation in the orbitofrontal cortex, medial PFC, insula, hippocampus, and parahippocampal gyrus in the post-meal condition (Holsen et al., 2006). Another study reported greater activation of the medial prefrontal cortex following a glucose challenge in subjects with Prader–Willi Syndrome compared to normal weight controls (Miller et al., 2007). Obese subjects compared to normal weight subjects exhibited greater activation of the medial prefrontal cortex, temporal, orbitofrontal cortex, caudate, and anterior cingulate when viewing food pictures after eating (Martin et al., 2009; Dimitropoulos et al., 2012). Interestingly, many of these regions (orbitofrontal cortex, insula, parahippocampal gyrus, anterior cingulate, and temporal lobe) exhibited greater BOLD signal following glucose compared to water in the insulin resistant group compared to the insulin sensitive group (Figs. 1 and 2, Table 4). The aforementioned studies did not examine whether

52

brain research 1558 (2014) 44–56

Table 5 – Correlation between insulin sensitivity and reactivity to food pictures during a glucose challenge. Region

Hem

k

tmax

MNI coordinates x

y

z

Negative correlation HC–C Inferior parietal lobule Posterior cingulate, precuneus Frontal superior medial, anterior cingulate, medial frontal gyrus Precentral gyrus Middle frontal gyrus Frontal mid/sup Mid temporal, middle temporal gyrus Frontal mid, superior frontal gyrus Precuneus Precentral Medial frontal gyrus

R R R R L R R L R L R

32 204 208 45 106 109 84 60 118 27 67

 6.76  6.64  5.92  5.89  5.76  5.63  5.54  5.39  5.32  4.77  4.63

48 8 24 64  24 24 58  20 26  58 12

 24  42 46 2 40 26 6 34  52 0 26

28 10 4 22 4 36 20 36 24 28 12

LC–C Precuneus, limbic lobe, posterior cingulate Superior frontal gyrus, frontal superior Superior temporal gyrus, mid temporal Rolandic operculum, parietal lobe Mid temporal

R R R R R

88 38 26 40 23

 7.32  5.39  5.07  4.70  4.66

10 22 52 48 52

 46 40  50  12 6

10 28 14 23 16

HC–LC Left Cerebrum Frontal lobe, anterior cingulate Frontal lobe Frontal lobe Frontal lobe

L R R R L

39 52 117 39 41

 5.30  5.21  5.00  4.91  4.68

 22 20 22 22  28

32 34 38 6 38

10 18 2 34 4

Positive correlation HC–LC Putamen Lentiform nucleus, putamen, pallidum

R R

71 33

5.40 4.17

32 20

4 2

8 6

Abbreviations: Hem ¼ Hemisphere, k ¼number of voxels within a cluster, HC ¼high calorie, LC¼ low calorie, C¼ control.

differences in brain responses seen in normal weight and obese subjects could be attributed to differences in insulin sensitivity. Because BMI was significantly higher in the insulin resistant group, we determined the extent to which BMI contributed to the results by modeling BMI both as a regressor and as a nuisance covariate. In contradistinction to the many regions that showed a correlation between 2 h G:I and BOLD signal to the various contrasts during a glucose challenge (Table 5), a correlation (positive) between BMI and activation was limited to the fusiform (HC–C contrast only). Furthermore, regions that showed a correlation between 2 h G:I and BOLD signal to the various contrasts during a glucose challenge remained when BMI was included as a nuisance covariate (results not shown). Similar findings were seen when waist circumference was substituted for BMI. Therefore, our results indicate that insulin sensitivity, rather than BMI or waist circumference (which differed between groups), offers the best explanation for BOLD response differences between groups. It is imperative that future studies determine the extent to which adiposity affects neural responses to food cues independent of insulin signaling in the brain.

3.2.

Strengths and limitations

A major strength of our study is the inclusion of a separate control scan rather than a baseline scan before each glucose challenge. This study design controls for potential habituation to food cues. The two scans were conducted in a counterbalanced fashion eliminating the potential of an order effect. Ingestion of water during the control scan controlled for feedback effects related to stomach distention. There also are significant limitations. First, insulin sensitivity was based on the 2 h G:I and HOMA2 derived from an OGTT. These proxies of insulin sensitivity have recognized limitations related to accuracy and reproducibility (Muniyappa et al., 2008). We attempted to address these concerns by requiring that subjects have both a 2 h G:Io1.5 and a HOMA2 o60% in order to be classified as insulin resistant. In addition, the OGTT for subject classification was conducted after an overnight fast, whereas the glucose challenge during scanning sessions was given after six hours of food abstinence. This was done so that subjects would not be inordinately hungry when scanned since this could potentially obscure group differences due to a ceiling effect. Conducting scans at midday rather than in the morning allowed us to

brain research 1558 (2014) 44–56

present a wider range of food pictures that would be considered appropriate for the time of day. While it is likely that insulin and glucose concentrations achieved during the OGTT were not identical to levels achieved by the glucose challenge administered during the scan, we assume that the relative concentrations of insulin and glucose during the OGTT were reflective of those present during the time of the scan. Second, the significance of our observed association between insulin sensitivity and appetitive brain responses on appetite as measured by ingestive behavior remains to be determined. The ingestion of glucose had a limited effect on satiation as measured by a post-scan hunger questionnaire. The accuracy of hunger ratings is questionable as subjects may have been influenced by preconceptions about how viewing food pictures or drinking a sweetened beverage affects hunger. This bias may be particularly confounding if met with acceptance by some subjects and resistance by others. A standardized meal may be superior to a glucose challenge when studying the neural correlates of satiety. However, because our primary objective was to determine the effect of insulin sensitivity on brain reactivity to food pictures, a glucose challenge rather than a standardized meal was a logical choice given its simplicity and specificity to address this question. Third, PCOS is a multifaceted condition characterized by oligo-ovulation or anovulation, hyperandrogenism, and polycystic ovaries. Potential confounding effects of ovarian steroids were minimized by imaging oligo-ovulatory subjects in the follicular phase. Although total testosterone levels were similar in the two groups, free testosterone was not determined. Therefore, we must temper our conclusion that these endocrine parameters were adequately controlled and did not contribute to the differential response that we attribute to insulin sensitivity. It will be important to determine if other forms of insulin resistance such as type 2 diabetes have a similar impact on brain reactivity to food cues.

3.3.

53

ovulatory dysfunction (PCOS Consensus Workshop Group, 2004). Excess androgen was defined as clinical evidence of hirsutism or acne on exam and/or elevated serum testosterone. Ovulatory dysfunction was defined as evidence of oligoovulation (o8 menstrual cycles/year) or secondary amenorrhea. The diagnosis of PCOS was made after excluding other etiologies of excess androgen or ovulatory dysfunction such as congenital adrenal hyperplasia, androgen secreting tumors, hyperprolactinemia, thyroid dysfunction, or Cushing's Syndrome. Exclusion criteria for imaging purposes included left-handedness, vegetarianism, claustrophobia, smoking, current or past drug or alcohol abuse, psychoactive medication, history of head injury, or presence of ferromagnetic material in the body. A total of twenty subjects completed the study. One subject was excluded from the analysis upon being diagnosis with type 2 diabetes.

4.2.

Calculating insulin sensitivity

The 2 h glucose to insulin ratio (2 h G:I) during an OGTT and the homeostasis model assessment of insulin sensitivity (HOMA2) were used as metrics of insulin sensitivity. Subjects drank 300 ml water containing 75 g d-glucose (ratio-GLUCOSE; ratiopharm) after an overnight fast. Blood was drawn immediately before and 2 h later. Insulin was determined using a monoclonal immunoenzymatic assay (Ultrasensitive Insulin Beckman Coulter UniCel DxI 800 Accesss Immunoassay System). Glucose was measured by the glucose oxidase method using the Beckman Coulter UniCels DxC 800 SYNCHRONs Clinical System. The glucose (mg/dl) and insulin (μU/ml) concentrations at two hours were used to calculate 2 h G:I. Fasting glucose and fasting insulin concentrations were inputted into the HOMA Calculator v2.2.2 downloaded from http://www.dtu.ox.ac.uk to calculate HOMA2. Subjects with a 2 h G:I o1.5 (Legro et al., 2004) and a HOMA2 sensitivity o60% (Marsh et al., 2013; Schofl et al., 2002) were designated insulin resistant.

Summary and conclusions 4.3.

Brain reactivity to food pictures was reduced during a glucose challenge in insulin sensitive but not insulin resistant PCOS patients, whereas brain reactivity was increased by a glucose challenge in insulin resistant but not insulin sensitive subjects. This opposite pattern of activity led to a significant interaction between sensitivity and condition and a negative correlation between activity during a glucose challenge and insulin sensitivity. These results suggest that insulin-induced inhibition of brain reactivity to food cues is compromised in the presence of peripheral insulin resistance. The extent to which dysfunctional insulin signaling leads to non-homeostatic eating and whether improving insulin sensitivity would alter brain reactivity and change ingestive behavior remains to be determined.

4.

Experimental procedures

4.1.

Subjects

The definition of PCOS for patient recruitment included clinical or biochemical evidence of excess androgens and

Experimental protocol

Subjects were scanned once during a glucose challenge and once during a water control. Subjects arrived at the imaging facility between 1400 and 1500 h having fasted since eating a normal breakfast between 0700 and 0900 h. Upon arrival at the facility, participants described what they had eaten for breakfast and confirmed that they had not eaten since breakfast. They filled out a safety checklist and a 9 item questionnaire that assessed hunger and wellness (i.e. physical and mental state indicated by alertness, warmth, anxiety, thirst, dizziness, and contentment). Responses were recorded on a Likert scale from 1 to 10 where 1 represented “not at all” and 10 represented extremely. On the first visit, subjects drank a 300 ml solution containing 75 g d-glucose (ratioGLUCOSE; ratiopharm) or an equivalent volume of water in 2 min or less (order was counterbalanced across all subjects). Subjects remained seated for five minutes before they were positioned in a 3.0 T whole body scanner (Siemens Magnetom Trio Tim syngo MR B15, Erlagen, Germany) equipped with a 12 channel head coil. After a three plane localizer scan, highresolution T1-weighted 3D structural images were acquired

54

brain research 1558 (2014) 44–56

(MPRAGE, TR 1760 ms; TE 2.2 ms; flip angle of 91; FOV 256 mm  256 mm; voxel size¼ 1  1  1 mm3). Thereafter, functional images were collected beginning 15 min after consuming glucose or water. Forty-four contiguous slices in the axial plane were collected using a T2- weighted echo planar imaging sequence (TR 3000 ms, TE 30 ms, flip angle¼ 841; FOV 240 mm  240 mm, 80  80 matrix; voxel size¼ 3  3  3 mm3, no gap). During the functional scans, pictures of high calorie (HC), low calorie (LC), and control (C) pictures were projected onto a screen at the head of the scanner bore and viewed through a mirror mounted on the head coil. The HC stimulus included pictures of savory, sweet, and high fat foods such as hamburgers, french fries, pizza, ice cream, and cake. The LC stimulus included pictures of water and fiber-based foods such as fresh vegetables, fruits, and salads. The mean caloric density (per 100 g) using the USDA National Nutrient Database (http://ndb.nal.usda. gov/) was significantly greater for HC compared to LC foods (2867113 (SD) and 45734 respectively; po0.0001). Control stimuli included pictures of landscapes, buildings, furniture, or recognizable household objects not associated with food. Photoshop was used to adjust the resolution of all images to 300 dpi and orient images in the landscape mode. Four pictures from the same category were presented in 20 s blocks (4.5 s for each picture separated by a fixation cross for 0.5 s). Blocks were presented in pseudo-random order to avoid the same picture category appearing in consecutive blocks. Blocks were separated by a fixation cross presented for 10 s. Sixty-four pictures in each category were presented in 16 blocks spread over 4 runs. Each run lasted 370 s, and runs were separated by approximately 30 s of rest. Subjects were instructed to think about how pleasant it would be to eat the food item shown. No specific instructions were given with regards to control pictures. In addition, they were told to focus their attention because a recall test would be given after the scan. Upon exiting the scanner, subjects completed a recall test consisting of 4 novel pictures and 4 repeat pictures in each of the 3 categories (24 pictures in total) administered using the quiz feature in WebCt and the 9 item questionnaire to assess post-scan hunger and wellness. All aspects of the study were repeated with the alternative condition 7–14 days later in twelve subjects with extreme oligoovulation or amenorrhea or 25–72 days in seven subjects, either because of scheduling difficulties (n¼4) or to ensure that both scans were conducted during the mid-follicular phase (between day 5 and 10 of different menstrual cycles) in subjects with semi-regular periods (n¼3). These 7 subjects were evenly distributed between insulin sensitive (n¼4) and insulin resistant groups. All procedures were reviewed and approved by the Queen's University Health Sciences and Affiliated Teaching Hospitals Research Ethics Board. The effect of insulin sensitivity on brain responses to food images following water consumption was described for a region of interest analysis applied to a subset of patients (Van Vugt et al., 2013).

London, UK) in MATLAB v.7.2 2006 software (Mathworks, Natick, MA). After discarding the first three images from each run in order to reduce non-steady state magnetic effects, all remaining functional images were aligned to the first functional image. The resulting mean functional image was normalized to a Montreal Neurological Institute (MNI) template brain (ICBM152). Images were re-sampled to 2-mm isotropic voxels and spatially smoothed with a 6-mm FWHM Gaussian kernel. Low frequency noise and signal drift were removed by applying a 128-second high-pass filter. Functional MRI data were analyzed using the General Linear Model approach (Worsley and Friston, 1995). Brain activation was modeled with a boxcar function convolved with the canonical hemodynamic response function. The goodness of fit of each stimulus-specific regressor was assessed for the BOLD time series in each voxel. Statistical parametric maps were generated for the contrasts HC–C, LC–C, HC–LC, and Food (HCþLC)–C for each subject (level one; fixed effects) and each condition (water or glucose). The individual contrast images were incorporated into a second level analysis (random effects) using the full factorial design approach of SPM8. This analysis consisted of a 2 (within subject condition: water or glucose)  2 (insulin sensitivity: sensitive or resistant) repeated measures ANOVA in order to determine the effect of condition, insulin sensitivity, and their interaction on BOLD responses to food pictures. The impact of insulin sensitivity on brain activation following glucose ingestion was examined by entering each subject's 2 h G:I value as a covariate in a one sample t-test. Body mass index (BMI) was included as a nuisance covariate to control for potential confounding effects of adiposity. The potential impact of adiposity on brain activation following glucose ingestion was similarly examined by including BMI as a covariate in a one sample t-test.

4.5.

Statistics

Based on a Monte Carlo simulation consisting of 1000 iterations applied to a functional image matrix of 80  80  44 and a FWHM Gaussian kernel of 6 mm, a voxel threshold of po0.001 (uncorrected) and a cluster extent threshold of 20 contiguous re-sampled voxels achieved a whole brain corrected p value of o0.05 (Forman et al., 1995; Slotnick et al., 2003). Signal intensity was calculated from parameter estimates for a 5 mm radius sphere centered on the peak voxel. Group means were compared by one way ANOVA and a Tukey–Kramer multiple comparison test using GraphPad Instat version 3.06. (La Jolla, CA). Group comparison of clinical and biochemical measurements were analyzed by two tailed t-tests. Within group comparisons of pre and post hunger scores were analyzed by Friedman Test (Nonparametric Repeated Measures ANOVA). Between group comparisons of hunger scores were analyzed by Kruskal–Wallis Test.

Acknowledgments 4.4.

Image pre-processing and analysis

Images were preprocessed and analyzed using Statistical Parametric Mapping 8 (SPM8; Wellcome Dept. Imaging Neuroscience,

The authors wish to acknowledge financial support from the Physicians Services Incorporated, the Botterell Foundation, and the Canadian Foundation for Women's Health.

brain research 1558 (2014) 44–56

r e f e r e n c e s

Anthony, K., Reed, L.J., Dunn, J.T., Bingham, E., Hopkins, D., Marsden, P.K., Amiel, S.A., 2006. Attenuation of insulin-evoked responses in brain networks controlling appetite and reward in insulin resistance: the cerebral basis for impaired control of food intake in metabolic syndrome? Diabetes 55, 2986–2992. Banks, W.A., Jaspan, J.B., Kastin, A.J., 1997. Selective, physiological transport of insulin across the blood-brain barrier: novel demonstration by species-specific radioimmunoassays. Peptides 18, 1257–1262. Baura, G.D., Foster, D.M., Porte, D., Kahn, S.E., Bergman, R.N., Cobelli, C., Schwartz, M.W., 1993. Saturable transport of insulin from plasma into the central nervous system of dogs in vivo. A mechanism for regulated insulin delivery to the brain. J. Clin. Invest. 92, 1824–1830. Benedict, C., Kern, W., Schultes, B., Born, J., Hallschmid, M., 2008. Differential sensitivity of men and women to anorexigenic and memory-improving effects of intranasal insulin. J. Clin. Endocrinol. Metab. 93, 1339–1344. Benoit, S.C., Clegg, D.J., Seeley, R.J., Woods, S.C., 2004. Insulin and leptin as adiposity signals. Recent Prog. Horm. Res. 59, 267–285. Bruning, J.C., Gautam, D., Burks, D.J., Gillette, J., Schubert, M., Orban, P.C., Klein, R., Krone, W., Muller-Wieland, D., Kahn, C. R., 2000. Role of brain insulin receptor in control of body weight and reproduction. Science 289, 2122–2125. Clegg, D.J., Riedy, C.A., Smith, K.A.B., Benoit, S.C., Woods, S.C., 2003. Differential sensitivity to central leptin and insulin in male and female rats. Diabetes 52, 682–687. Del Parigi, A., Gautier, J.F., Chen, K., Salbe, A.D., Ravussin, E., Reiman, E., Tataranni, P.A., 2002. Neuroimaging and obesity: mapping the brain responses to hunger and satiation in humans using positron emission tomography. Ann. N. Y. Acad. Sci. 967, 389–397. Dimitropoulos, A., Tkach, J., Ho, A., Kennedy, J., 2012. Greater corticolimbic activation to high-calorie food cues after eating in obese vs. normal-weight adults. Appetite 58, 303–312. Dunaif, A., Segal, K.R., Futterweit, W., Dobrjansky, A., 1989. Profound peripheral insulin resistance, independent of obesity, in polycystic ovary syndrome. Diabetes 38, 1165–1174. Dunaif, A., 1997. Insulin resistance and the polycystic ovary syndrome: mechanism and implications for pathogenesis. Endocr. Rev. 18, 774–800. Forman, S.D., Cohen, J.D., Fitzgerald, M., Eddy, W.F., Mintun, M.A., Noll, D.C., 1995. Improved assessment of significant activation in functional magnetic resonance imaging (fMRI): use of a cluster-size threshold. Magn. Reson. Med. 33, 636–647. Foster, L.A., Ames, N.K., Emery, R.S., 1991. Food intake and serum insulin responses to intraventricular infusions of insulin and IGF-I. Physiol. Behav. 50, 745–749. Guthoff, M., Stingl, K.T., Tschritter, O., Rogic, M., Heni, M., Stingl, K., Hallschmid, M., Haring, H.U., Fritsche, A., Preissl, H., Hennige, A.M., 2011. The insulin-mediated modulation of visually evoked magnetic fields is reduced in obese subjects. PLoS ONE 6, e19482. Hallschmid, M., Higgs, S., Thienel, M., Ott, V., Lehnert, H., 2012. Postprandial administration of intranasal insulin intensifies satiety and reduces intake of palatable snacks in women. Diabetes 61, 782–789. Hinton, E.C., Parkinson, J.A., Holland, A.J., Arana, F.S., Roberts, C., Owen, A.M., 2004. Neural contributions to the motivational control of appetite in humans. Eur. J. Neurosci. 20, 1411–1418. Holsen, L.M., Zarcone, J.R., Brooks, W.M., Butler, M.G., Thompson, T.I., Ahluwalia, J.S., Nollen, N.L., Savage, C.R., 2006. Neural mechanisms underlying hyperphagia in Prader–Willi Syndrome. Obes. Res. 14, 1028–1037.

55

Holsen, L.M., Zarcone, J.R., Thompson, T.I., Brooks, W.M., Anderson, M.F., Ahluwalia, J.S., Nollen, N.L., Savage, C.R., 2005. Neural mechanisms underlying food motivation in children and adolescents. NeuroImage 27, 669–676. Killgore, W.D.S., Young, A.D., Femia, L.A., Bogorodzki, P., Rogowska, J., Yurgelun-Todd, D.A., 2003. Cortical and limbic activation during viewing of high- versus low-calorie foods. NeuroImage 19, 1381–1394. Kroemer, N.B., Krebs, L., Kobiella, A., Grimm, O., Vollstadt-Klein, S., Wolfensteller, U., Kling, R., Bidlingmaier, M., Zimmermann, U.S., Smolka, M.N., 2013. (Still) longing for food: insulin reactivity modulates response to food pictures. Hum. Brain Mapp. 34, 2367–2380. LaBar, K.S., Gitelman, D.R., Parrish, T.B., Kim, Y.H., Nobre, A.C., Mesulam, M.M., 2001. Hunger selectively modulates corticolimbic activation to food stimuli in humans. Behav. Neurosci. 115, 493–500. le Roux, C.W., Patterson, M., Vincent, R.P., Hunt, C., Ghatei, M.A., Bloom, S.R., 2005. Postprandial plasma ghrelin is suppressed proportional to meal calorie content in normal-weight but not obese subjects. J. Clin. Endocrinol. Metab. 90, 1068–1071. Legro, R.S., Castracane, V.D., Kauffman, R.P., 2004. Detecting insulin resistance in polycystic ovary syndrome: purposes and pitfalls. Obstet. Gynecol. Surv. 59, 141–154. Lo, J.C., Feigenbaum, S.L., Yang, J., Pressman, A.R., Selby, J.V., Go, A.S., 2006. Epidemiology and adverse cardiovascular risk profile of diagnosed polycystic ovary syndrome. J. Clin. Endocrinol. Metab. 91, 1357–1363. Malik, S., McGlone, F., Bedrossian, D., Dagher, A., 2008. Ghrelin modulates brain activity in areas that control appetitive behavior. Cell Metab. 7, 400–409. Marsh, C., Berent-Spillson, A., Love, T., Persad, C.C., Pop-Busui, R., Zubieta, J.K., Smith, Y.R., 2013. Functional neuroimaging of emotional processing in women with polycystic ovary syndrome: a case-control pilot study. Fertil. Steril. 100, 200–207. Martin, L.E., Holsen, L.M., Chambers, R.J., Bruce, A.S., Brooks, W. M., Zarcone, J.R., Butler, M.G., Savage, C.R., 2009. Neural mechanisms associated with food motivation in obese and healthy weight adults. Obes. Res. 18, 254–260. Miller, J.L., James, G.A., Goldstone, A.P., Couch, J.A., He, G., Driscoll, D.J., Liu, Y., 2007. Enhanced activation of reward mediating prefrontal regions in response to food stimuli in PraderΓC¸oˆWilli syndrome. J. Neurol., Neurosurg. Psychiatry 78, 615–619. Muniyappa, R., Lee, S., Chen, H., Quon, M.J., 2008. Current approaches for assessing insulin sensitivity and resistance in vivo: advantages, limitations, and appropriate usage. Am. J. Physiol. – Endocrinol. Metab. 294, E15–E26. Obici, S., Feng, Z., Karkanias, G., Baskin, D.G., Rossetti, L., 2002. Decreasing hypothalamic insulin receptors causes hyperphagia and insulin resistance in rats. Nat. Neurosci. 5, 566–572. Page, K.A., Seo, D., Belfort-DeAguiar, R., Lacadie, C., Dzuira, J., Naik, S., Amarnath, S., Constable, R.T., Sherwin, R.S., Sinha, R., 2011. Circulating glucose levels modulate neural control of desire for high-calorie foods in humans. J. Clin. Invest. 121, 4161–4169. Pagotto, U., 2009. Where does insulin resistance start? Diabetes Care 32, S174–S177. Paik, K.H., Choe, Y.H., Park, W.H., Oh, Y.J., Kim, A.H., Chu, S.H., Kim, S.W., Kwon, E.K., Han, S.J., Shon, W.Y., Jin, D.K., 2006. Suppression of acylated ghrelin during oral glucose tolerance test is correlated with whole-body insulin sensitivity in children with Prader–Willi Syndrome. J. Clin. Endocrinol. Metab. 91, 1876–1881. PCOS Consensus Workshop Group, 2004. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome (PCOS). Hum. Reprod. 19, 41–47.

56

brain research 1558 (2014) 44–56

Plum, L., Schubert, M., Bruning, J.C., 2005. The role of insulin receptor signaling in the brain. Trends Endocrinol. Metab. 16, 59–65. Porubska, K., Veit, R., Preissl, H., Fritsche, A., Birbaumer, N., 2006. Subjective feeling of appetite modulates brain activity: an fMRI study. NeuroImage 32, 1273–1280. Schofl, C., Horn, R., Schill, T., Schlosser, H.W., Muller, M.J., Brabant, G., 2002. Circulating ghrelin levels in patients with polycystic ovary syndrome. J. Clin. Endocrinol. Metab. 87, 4607–4610. Schulingkamp, R.J., Pagano, T.C., Hung, D., Raffa, R.B., 2000. Insulin receptors and insulin action in the brain: review and clinical implications. Neurosci. Biobehav. Rev. 24, 855–872. Slotnick, S.D., Moo, L.R., Segal, J.B., Hart, J., 2003. Distinct prefrontal cortex activity associated with item memory and source memory for visual shapes. Cogn. Brain Res. 17, 75–82. St-Onge, M.P., Sy, M., Heymsfield, S.B., Hirsch, J., 2005. Human cortical specialization for food: a functional magnetic resonance imaging investigation. J. Nutr. 135, 1014–1018. St-Pierre, D.H., Karelis, A.D., Coderre, L., Malita, F., Fontaine, J., Mignault, D., Brochu, M., Bastard, J.P., Cianflone, K., Doucet, E., Imbeault, P., Rabasa-Lhoret, R., 2007. Association of acylated and nonacylated ghrelin with insulin sensitivity in overweight and obese postmenopausal women. J. Clin. Endocrinol. Metab. 92, 264–269. Strubbe, J.H., Porte, J., Woods, S.C., 1988. Insulin responses and glucose levels in plasma and cerebrospinal fluid during fasting and refeeding in the rat. Physiol. Behav. 44, 205–208.

Tschritter, O., Preissl, H., Hennige, A.M., Stumvoll, M., Porubska, K., Frost, R., Marx, H., Klosel, B., Lutzenberger, W., Birbaumer, N., Haring, H.U., Fritsche, A., 2006. The cerebrocortical response to hyperinsulinemia is reduced in overweight humans: a magnetoencephalographic study. PNAS 103, 12103–12108. van der Laan, L.N., de Ridder, D.T.D., Viergever, M.A., Smeets, P.A. M., 2011. The first taste is always with the eyes: a metaanalysis on the neural correlates of processing visual food cues. NeuroImage 55, 296–303. Van Vugt, D.A., Krzemien, A.A., Alsaadi, H., Palerme, S., Reid, R.L., 2013. Effect of insulin sensitivity on corticolimbic responses to food picture in women with polycystic ovary syndrome. Obes. Res. 21, 1215–1222. Van Vugt, D.A., 2010. Brain imaging studies of appetite in the context of obesity and the menstrual cycle. Hum. Reprod. Update 16, 276–292. Vidarsdottir, S., Smeets, P.A.M., Eichelsheim, D.L., van Osch, M.J. P., Viergever, M.A., Romijn, J.A., van der Grond, J., Pijl, H., 2007. Glucose ingestion fails to inhibit hypothalamic neuronal activity in patients with type 2 diabetes. Diabetes 56, 2547–2550. Woods, S.C., Lotter, E.C., McKay, L.D., Porte, D., 1979. Chronic intracerebroventricular infusion of insulin reduces food intake and body weight of baboons. Nature 282, 503–505. Worsley, K.J., Friston, K.J., 1995. Analysis of fMRI time-series revisited again. NeuroImage 2, 173–181.