Food and Mood: Exploring the determinants of food choices and the effects of food consumption on mood among women in Inner London

Introduction: The aim of this study was to explore the relationship between food and mood against the backdrop of increased mental health and nutrition cognizance within public health and scientific discourses. Mood was defined as encompassing positive or negative affect. 
Methodology: A constructionist qualitative approach underpinned this study. Convenience sampling in two faith-based settings was utilised for recruiting participants, who were aged 19-80 (median,48) years. In total 22 Christian women were included in the research, eighteen were in focus groups and four were in individual semi structured interviews. All were church-attending women in inner London. A thematic analysis was carried out, resulting in four central themes relating to food choice and food-induced mood states. 
Findings: Women identified a number of internal and external factors as influencing their food choices and the effect of food intake on their moods. Food choice was influenced by mood; mood was influenced by food choice. Low mood was associated with unhealthy food consumption, apparent addiction to certain foods and overeating. Improved mood was associated with more healthy eating and eating in social and familial settings. 
Discussion: Findings indicate food and mood are interconnected through a complex web of factors, as women respond to individual, environmental, cultural and social cues. Targeting socio-cultural and environmental influences and developing supportive public health services, via faith-based or community-based institutions could help to support more women in their struggle to manage the food and mood continuum. Successful implementation of health policies that recognise the psychological and social determinants of food choice and the effect of food consumption on mood, is essential, as is as more research into life-cycle causal factors linking food choice to mood.


Introduction
The connection between food, nutrients and brain function is a developing area of interest within nutrition and mental health research (Sarris et al., 2015). While neural mechanisms underpinning connections between mood and diet are not fully understood, research indicates that poor diets are risk factors for depression and other psychological disorders (Akbaraly et al., 2009), whereas healthy eating habits have been linked to mental health benefits (Conner et al., 2017). Early scientific evidence identified a link between nutrition and brain function (Biggio et al., 1974;Fernstrom and Wurtman, 1974), and the presence of strong associations between nutritional quality and mental health (Sarris et al., 2015).
Mood and emotions are distinct from one another and yet interconnected, as denoted in a range of theoretical perspectives (Beedie et al., 2005). For the purposes of this research, 'mood' provides an umbrella concept that captures "the two primary dimensions of mood-Positive and Negative Affect" (Watson et al., 1988. The psychological determinants of food choice haven been linked to both transitory and protracted mood states, while dietary intake of specific nutrients are understood to influence biological processes relating to cognition, emotion and behaviour. Some studies suggest that mood-state is both a precursor and an outcome of food experiences (Gibson, 2006). Thus, determining beginning and end points of the temporal food and mood cycle can be difficult to distinguish, though it has been suggested that foods may instigate the cycle (Hendy et al., 2012;White et al., 2013).

The effect of mood on dietary intake
Basic biological and physiological needs fuel food intake through provision of calories and nutrients needed to function and to balance hunger and satiety signals, yet a range of factors are implicated in food choice. Studies have examined the link between food choice and complex neurocircuitry, implicating hormonal mechanisms, chronic stress, cognitive load, and hedonic sensory processes (Schellekens et al., 2012;Klatzkin et al., 2018;Shiv and Fedorikhin, 1999;Ward & Mann, 2000;Moore and Bovell, 2008). Dietary choice may be influenced by positive or negative mood states, including complex internal individual cues which may signify associations of reward and deprivation (Gardner et al., 2014;Singh, 2014). External factors including social and environmental opportunities may perpetuate these cues. Positive affect has been linked to social experiences of eating while isolated consumption has been associated with loneliness and comfort eating (Oxford Economics and the National Centre for Social Research, 2018;Locher et al., 2005). Low mood is also closely linked to low energy and tiredness (Garrosa et al., 2008), prompting food selections to provide 'highs', thereby implicating cravings for foods that stimulate mood activating dopamine neurons of addiction and reward (De Macedo et al., 2016). Active pleasure-seeking behaviours to reduce negative states (Freud, 1920) or avoidance distraction (Spoor et al., 2017) can fuel unconscious and conscious eating patterns to mitigate negative affect, and often involve sweet or fatty products (Evers et al., 2010;Tomiyama et al., 2011). Innate preferences for sweet taste (Keskitalo et al., 2007) and palatable foods (Yeomans et al., 2004) are reinforced by childhood and learned experiences (De Cosmi et al., 2017), while social constructions and marketing forces promoting food as 'reward', 'comforting' 'stress relieving', and 'indulgent' can strengthen these emotional signals (Locher et al., 2005). Using foods high in fats and sugars as coping strategies is linked to detrimental psychological and physiological effects, and shortterm mood 'boosts' can result in subsequent prolonged low-mood state (Freeman and Gil, 2004;Kiecolt-Glaser, 2010).

The effect of dietary intake on mood
Neurological activity linked to mood and behaviour has been associated with nutrients and dietary intake (Jacka et al., 2017;Strang et al., 2017). A complex interplay of brain and gut activity are understood to act as pathways for potential effects on mood, cognitive and emotional processes (Zagon, 2001), including tryptophan and the gut microbiome, excess sugar and fat consumption, micronutrient function and inflammatory processes (Lomagno et al., 2014;Jenkins et al., 2016;Kroes et al., 2014;Liao et al., 2018). Sugar and fat ingestion may temporarily stimulate the endogenous μopioid receptor system and dopamine pathways to alleviate stress and negative mood state (Tuulari et al., 2017;Wenzel and Cheer, 2018). Instantaneous mood improvement from palatable food in the postprandial stage can forge repeated and habitual coping mechanisms for mood management (Macht and Mueller, 2007). Observational studies, cross-sectional data and longitudinal research point to an association between fruit and vegetable consumption and improved mental health outcomes (Rooney et al., 2013;Mujcic and Oswald, 2016). Links have been established between high sugar and fat consumption (Liao et al., 2018), low intake of omega-3 fatty acids and a depressed mood (Akbaraly et al., 2009), while Mediterranean pattern diets are associated with improved long-term mental health conditions through reductions in inflammation and oxidative stress (Sánchez-Villegas et al., 2009;Conner et al., 2017). In addition to improving long-term mental health, a wealth of studies has demonstrated improvements on short-term well-being and positive affect following fruit and vegetable consumption (Wahl et al., 2017;White et al., 2013;Blanchflower et al., 2013), while some trials support a causal association between well-being and fruit and vegetable consumption (Conner et al., 2017). As such, both short-term and long-term mood effects are significant in the food-mood relationship.

Women and the food-mood relationship
Distinctions have been drawn between men and women concerning nutrition and health (Kiefer et al., 2005), and notable gender disparities are present in the physiological and psychological processes of the food-mood connection (Arganini et al., 2012). Increased limbic system and prefrontal cortex activity have been linked to mood state and anxiety in women who are prone to elevated stress response and chronic stress vulnerability (Matud, 2004;McDonough and Walters, 2001). More women in England experience anxiety and lower moods compared with men, with documented rates of 19% and 12%, respectively (McManus et al., 2016). Social roles that women tend to occupy may be significant in this response (Oksuzyan et al., 2010;WHO, 2009). A higher proportion of women than men are likely to eat in response to stress, anxiety and low mood (Lafay et al., 2001), with a number of factors implicated in this process, including sleep deprivation (Saleh-Ghadimi et al., 2019), and socioeconomic factors (Thompson et al., 2018). Heightened cortisol response and resulting hypothalamic pituitary adrenal axis responsivity (Klatzkin et al., 2018) can lead to increased energy intake via sweet and fatty food products (Zellner et al., 2006;Roberts, 2008). Feelings of guilt in the post-consumption period may contribute to increased low mood which is particularly pronounced in women (Wansink et al., 2003), and amplified in the presence of body image ideals which has been observed in both younger and older women (Bedford and Johnson, 2006).

Faith, health beliefs and wellbeing
Health and health-related practices may shape the interplay between food and mood among individuals and populations (Wahl et al., 2017). Religious beliefs in particular may be an important aspect of the food-mood connection within faith communities, due to the widely held conviction of the interconnectedness between mind, body and spirit (Koenig, 2012).

The public health context: nutrition and mental health
The World Health Organisation (1946) positions mental well-being as a central aspect of its definition of health and as a core principle embedded in its constitution. However, few countries' health systems adequately address mental health and well-being (Lake and Turner, 2017). Depression is the largest contributor to global disability and poor health, affecting over 300 million people (WHO, 2019). In the UK approximately 19% of people suffer from common mental health problems (Office for National Statistics, 2019). Depressed persons are at risk of developing physical illnesses, and have a 58% increased risk of obesity (Luppino et al., 2010). In light of the growing evidence surrounding nutrition and mental health, this has important implications where highly processed foods and sugary snacks are readily available (Otter, 2012). Chronic exposure to such products perpetuates unhealthy eating patterns, which can also be linked to mood disorders (Breymeyer et al., 2016), as well as noncommunicable diseases linked to obesity, of which depression and anxiety, are known comorbidities (Luppino et al., 2010;Bjerkeset et al., 2008). Public health recommendations suggest a reduction of free sugars to <5% of daily calorie intake. However UK adult consumption exceeds this recommended limit by more than double (Scientific Advisory Committee on Nutrition, 2015;Public Health England, 2015), with factors such as bottle feeding and unhealthy maternal gut microbiome (Keith et al., 2019), marketing forces (Locher et al., 2005), and sweet taste preferences (Keskitalo et al., 2007) all influencing intake. In addition, inadequate intakes of DHA and EPA (available mainly from oily fish), and fruits and vegetables risk fatty acid, vitamin and mineral deficiencies needed for optimum brain function and mental health (British Nutrition Foundation, 2018). Age, socio-economic factors and familial structures all may affect consumption levels (Verbeke and Vackier, 2005).
Few studies have addressed women's perspectives to find meaning in the food and mood relationship from their point of view. The aim of this qualitative study was to explore key issues surrounding food and mood through the experiences of a sample of women with faith, in inner London. The research also attempted to understand the role faith may play in healthy eating practices and perceptions.

Study design
As informed by Geertz (1973), an adaptive research design underpinned this study to yield 'thick' description. Embedded in a constructivist paradigm whereby new data and meaning is continuously generated through interactions between the researcher and the researched (Bryman, 2016), a qualitative design was selected to allow the quotidian narratives of women to formulate the course of the enquiry. As such this study is underpinned by subjectivist ontological reasoning to elucidate meaning from the women's experiences, while deductive processes generated public health data and evidence-based research relevant to the topic.

Setting
Convenience sampling was used, taking advantage of respected gatekeeper access to the population. The study was carried out in two churches in inner London, where common mental health conditions account for a large burden of disease. Inner London anxiety levels in particular are noted to be higher than those in outer London and across the UK, with prevalence estimates at 42.1% (Authority, 2014).

Participants and sampling
The study sample comprised 22 women. Sample size was determined by the level of "information power" needed to yield diversity of voices within a set time-frame (Malterud et al., 2016). Study eligibility was defined by age and sex so that only women aged 18 or over were recruited and only those able to speak English and provide written consent. Convenience sampling was employed via one personal church context (Church 1) as an advantageous method that afforded access to a group of women in one setting within a set time-period (Marshall, 1996). One further church (Church 2) was contacted via email and chosen based on locality and accessibility. Sampling from Church 2 was pursued in order to diversify participants, interview women unfamiliar to the researcher, and to enable within-method triangulation of data. Participant ages ranged from 19 to 80 (mean=48) years. The sample contained mostly White British (36%), single (45.5%) women educated beyond secondary school (59%).

Data collection
Participants were recruited via posters displayed in church premises as a cost-effective and straightforward recruitment tool that is associated with a high response rate for research studies (Krusche et al., 2014). "Informational redundancy" informed the timing of the conclusion of data collection (Sandelowski, 2008, p875). Participants were interviewed through focus groups as the principle form of data collection and individual semi-structured interviews as a secondary method. Utilising both focus groups and individual interviews afforded the opportunity to mitigate weaknesses inherent in each method. Interviews were conducted within church grounds and recorded using an audio device following participant consent.

Focus Groups
Focus groups are noted as powerful tools for validating and empowering voices of women, and for enabling solution-focused sharing (Madriz, 2000). Focus groups were additionally chosen as a means of harnessing dynamism in collective experiences and to elicit varied opinions. Within focus groups some women were well-known to each other, and these pre-formed 'natural groups' facilitated flowing discourse that enhanced the women's narratives to provide rich data (Kitzinger, 1995). Seating was arranged in a circular fashion to ensure participants were equally positioned and to reduce the presence and power of the researcher, which afforded an equalising environment for openness and sharing (Kamberelis and Dimitriadis, 2010). Group size (3 focus groups of 6 participants each) and length (60-90 minutes) were planned in consonance with recommended qualitative health research recommendations (Green and Thorogood, 2018). After adjusting to suit participant needs, focus groups included 7, 4 and 7 participants per group.

One-to-one interviews
Semi-structured individual interviews were conducted with 4 participants to explore topics in greater depth, and as a validation strategy for methodological triangulation to increase depth in findings (Denzin, 2017;Thurmond, 2001). Interviews lasted between 30-35 minutes.

Topic guide
Questions were designed to draw out personal responses and formulated through analysis of extant research on women, health, nutrition, mood, and mental health. The topic guide included questions to explore "emotional cues, moods and feelings," that affect food choice, as suggested in Furst et al.'s (1996) conceptual model of food choice. The topic guide was piloted on 3 individuals to establish its feasibility as a research tool (Creswell, 2009), and questions were adjusted accordingly.

Data analysis
To uphold anonymity, names and personal details were removed and participants' responses are displayed using numbers as follows: those who were part of Focus Group A are presented as P1 to P7; those in Focus Group B -presented as P8 to P11 and individual interviewees as P12 to P15. All participants P1 to P15 attended Church 1 which was attended by the interviewer; participants in Focus Group C were labelled P16 to P22 and attended Church 2.
The three focus groups and four individual interview recordings were fully transcribed verbatim. Thematic analysis was carried out to elicit patterns and themes within the data (without the use of software). This method was not fixed to a particular theoretical framework, thereby allowing participant voices to better drive the enquiry (Braun and Clarke, 2006).
An identical data analysis process was followed for focus groups and individual interviews, however initial analyses were performed separately before drawing together, in order to trace distinguishing patterns captured in either method that may have produced different results. Further cross-analysis took place between Church 1 and Church 2 in order to compare responses.
Four important themes emerged within the data, with several subthemes, which are presented in Table 1.

Quality assurance procedures
Reference to the Critical Appraisal Skills Programme qualitative checklist (2017) enabled conscious reflexive practice throughout the process of collection and analysis. Data were shaped from interactions between the lead researcher and the participants, and as such necessitated a rigorous process of reflexive awareness. Co-authors reviewed interpretations and conclusions, with no areas of disagreement and full consensus reached. Seale (1999) suggests researcher beliefs can exist beyond consciousness, and in recognition that epistemological views and methodological assumptions cannot be fully detached from the researcher, and in order to be open to critique, transparency was sought in all aspects of the research process (Green and Thorogood, 2018).

Ethics
This study was carried out with full adherence to University of Westminster's ethical principles as outlined in the Code of Good Research Practice and Code of Practice Governing the Ethical Conduct of Research (2017a; 2017b; Ethics code: ETH1718-1274, Class 1). Verbal permission was obtained from church leaders to conduct research on their premises. Participants received information sheets via email which included an introduction to the study, description of interview procedures, participant's rights, and the researcher's contact details. To ensure a safe and ethical research environment (Krueger and Casey, 2014), participants were informed of confidentiality processes prior to taking part in the study, and were assured that they could withdraw at any stage with no consequences.

Findings 1 and Discussion
Objectives of this study were met through rich and descriptive material that highlighted psychological facets of the food-mood relationship as entwined with socio-cultural, environmental and physiological variables. Emergent patterns in analysis suggest pre-prandial and post-prandial mood state encompassed varying degrees of the positive and negative affect spectrum, with conscious and reflective mechanisms implicated. Data suggests healthy eating patterns were linked to positive affect, while excess sugar and fat consumption were linked to both positive and negative affect. Themes are illustrated below with excerpts from the interviews. 2

Theme 1: Healthy eating incentives and challenges
"Eating well gives me a real mood boost, it really does. I think there is a clear link between how we view our food and how we feel when we eat it. I feel much happier eating food I know is good for me…" (P10) Nearly all women in the sample described healthy eating as linked to health benefits. Healthy choices were often preceded by a good mood which supports studies that implicate temporal construal processes in determining food choices; positive affect has been linked to choices serving long-term goals such as health, while negative affect prompts greater likelihood of indulging in unhealthier foods (Gardner et al., 2014). Positive post-prandial effects were often experienced when preceded by good mood and alongside ongoing patterns of healthy eating.

"…I've started to change how I eat and I do feel happy …I feel satisfied when I eat my nice mixed salad with fish…" (P13)
Women frequently reported both improved energy and wellbeing when incorporating other wholefoods included in Mediterranean patterns of eating, such as fish, wholegrains and leafy greens. Omega-3 fatty acids and essential nutrients such as zinc, folate and B-vitamins found in such foods have been consistently linked to improved mental health and lower rates of depression ( , 2009;Sánchez-Villegas et al., 2009). This may be due to the synergistic effects of essential fatty acids, B vitamins and minerals in many of the women's healthier choices, as well as the presumed effects of nutrients and sense of 'virtue' in consuming such foods.
"I always try to go for like a piece of fruit first though sometimes it just doesn't work… I still eat the chocolate!...It is like a drug. I remember giving it up for Lent and being really moody" (P14) While healthy eating was associated with numerous benefits, many women felt drawn to unhealthy products, some for energy and a temporary boost, others for mood enhancement. Improved health was cited as a goal for some women who noted being 'on track' with healthy eating for a finite amount of time before re-engaging with unhealthy habits. Some described making conscious adjustments to consumption based on health concerns, while for others changing behaviours emerged as difficult, despite the desire to make nutritious choices.
"…when you go in into our office… there's a whole set of cabinets where we have our snacks, and it's like every day someone will bring in biscuits or cakes …you've just got this constant temptation" (P10) Hunger signalling was amplified through factors that inhibited the ability to align intention with behaviour, including behavioural cues that were deeply embedded into the women's everyday lives and environments. Everyday settings affected choices and quantity of food consumed, as women distinguished between eating patterns in work, social and home environments.
"Tiredness sometimes actually makes me eat more… yesterday I had Domino's pizza…I hate it but I so was tired, I felt ill with tiredness, so I ate it" (P9) When hungry, tired and pressed for time, women expressed increased likelihood of choosing less nutritious foods and more convenience products.
"…to eat more healthily…It takes a degree of preparation that isn't the same as just whacking something in the microwave" (P12) Time emerged as a precious commodity among the women, and was closely linked to hunger, environment and life stage. Working mothers contrasted with retired women or women who worked from home who were able to plan meals and expressed fewer pressures. For some, their roles as family providers affected energy levels, mood, food choices and this restricted their capacity to take care of themselves. In a position of low energy women expressed a finite amount of inner resources to make healthy choices, at times leading to post-prandial energy decrease. Such experiences are indicative of a competing demand between cognitive processes and affect when making decisions--of "heart and mind in conflict'" as explored by Shiv and Fedorikhin (1999). Greater levels of cognitive load may account for possible energy lost and overload of the prefrontal cortex, thereby inhibiting cognition and stimulating affective response among the women.

Theme 2: Emotional eating
Assuaging 'emotional hunger' emerged as significant for women in this sample. Food choice and eating behaviour was fuelled by a host of emotional drivers and underpinned by an overarching recognition that food was used for purposes beyond purely satiation of hunger. Eating behaviour was complex as women sought food in response to a variety of negative states, which has been observed in both healthy women and those with eating disorders (Spoor et al., 2017). Avoidance coping, stress-relief, comfort, pursuing emotional 'equilibrium,' and seeking distraction were key to women's experiences.

"If I'm very upset, I overeat. If it's been an emotional day … and I'm in a bad mood…I'll have chocolate, crisps, wine. McDonalds is a number 1 contender" (P17)
Overeating was linked to a dysphoric mood for many women, as corresponds with prior research (Evers et al., 2010;Wansink et al., 2003). Both high negative and low positive affect were implicated in this eating behaviour.

"I overeat sometimes if I'm a bit bored, so if I'm under occupied I might fill time…" (P14)
Overeating in low positive affect was linked to boredom and distracted eating, and rarely involved engagement with other persons. Social isolation is linked to poor mental health (Holt-Lunstad et al., 2015) and comfort eating to feelings of social isolation (Locher et al., 2005). It is therefore unsurprising that eating alone in the present study was connected to low mood and sadness.
"it's kind of like…a vicious cycle thing…when I'm in a low mood I'll eat for comfort, and then after it's like…my outlook would be more negative than usual, just a general grey feeling" (P15) Food consumption to relieve negative feelings sometimes perpetuated low mood state and stimulated cyclical 'coping' patterns and affect loops. Women reported low mood as a precursor to choosing convenience foods, as well as a consequence of consuming processed and sugary foods, which has been observed in other research (Freeman and Gil, 2004).
"If I'm very upset, I overeat…and that makes me very upset… I eat all this sugary stuff, it just drops me, making me really cranky" (P18) Fuelled by the ubiquitous belief of certain products providing 'comfort' (Wansink et al., 2003), some women sought mood melioration in eating. Foods consumed in low mood included sugary, fatty and processed items and low mood also stimulated overconsumption of these products. Meeting short-term need was thereby underpinned by eating patterns that were conducive to the immediate goal of relief and obstructive to the longer-term intention of feeling good.
Several women reported stress as influencing their eating habits which is consistent with research that links emotion-driven appetites in women to high-stress levels (Klatzkin et al., 2018). For women in this sample, stress-induced eating invoked impulsivity in food choice and inhibited portion control, indicating cognitive demand and depletion which has been linked to reduced ability to withstand indulgences (Fedorikhin and Patrick, 2010).

Theme 3: Relationships with food
Relationships with food were complex, multi-layered, and subject to individual variability. Individual value systems, interpersonal level factors as well as external cues and environments impacted food choice and post-prandial mood response. Life stage, childhood experiences, taste preferences, individual and familial health concerns, self-perceptions, attitudes and beliefs were influential in colouring relationships with food. Food and eating represented enjoyment and frustration, anxiety and contentment, love and hate, comfort and reward, a means to nurture bodies, children and relationships, and a mechanism by which women asserted control over their lives.
"…as a woman, being able to provide food others may enjoy and share, while different to the necessity aspect, can be strong" (P4) Age, life stage and familial circumstances varied throughout the sample. While aspects of the women's lifestyles were different, including students living outside of familial home environments, mothers who worked, and retired women, descriptions of the women's personal roles as 'providers' were woven into the narratives, and at other times amid life pressures, and at others as a source of enjoyment and fulfilment. Adapting eating behaviours to favour the needs of family and cooking for others as an expression of love was important to some of the women and enhanced their relationships with food. Eating in the context of family and primary relationships added pleasure and value to the women's experiences and was linked to positive affect, which is consistent with research that links eating with others to with improved mental health and wellbeing (Oxford Economics and the National Centre for Social Research, 2018).
"…it's like a celebration. There's something about food and feasting and it makes people happy doesn't it... eating food with friends is very sociable, very nice, you know, it's a joyous activity" (P14) Social eating was frequently viewed as a celebration, with food a central feature of togetherness and sharing. Consumption behaviours sometimes changed to include bigger portions and foods that were not necessarily consumed at home, yet overeating in social circumstances was not linked to pre-or post-consumption negative affect, in contrast to overconsumption during isolated eating, which was driven by negative mood and prolonged negative affect.
While links have been established between negative mood and unhealthy eating, studies conflict as to whether positive mood stimulates such eating patterns (Bongers et al., 2013;Fedorikhin and Patrick, 2010). Findings from the present study suggest positive mood state is linked to both self-regulation as well as overindulgence. Good mood preceded and followed nutritious food choices, while good mood in social contexts preceded and followed both healthy and less healthy eating behaviours.
Ideas of eating as joined to togetherness and relationships were reinforced through some of the women's childhood experiences. Childhood foods prompted high emotional valence, and women who drew happiness from specific foods often had positive connections with these foods through familial experiences. Women with negative childhood experiences of eating specific foods expressed barriers to consuming such foods. Preferences and food aversions develop throughout the life course (De Cosmi et al., 2017), and some women were aware their habits and preferences were forged through childhood eating experiences, such as eating larger portions or specific foods. Women with negative experiences consequently drew associations with certain tastes, smells and textures, and expressed strong aversions to both certain healthy and unhealthy foods. Foods that provided contentment for women with positive experiences included traditional home-made dishes which held particular resonance as part of their personal associations. Mood-boosting properties of these foods may be attributed in part to an emotional association, however home-made dishes often contained nutritional benefits, such as home-cooked curries. Aspects of love and happiness associated with familial dishes may be enhanced by the mechanistic effects of nutrients. For example, essential amino acids such as tryptophan present in protein-based dishes act as precursors to serotonin, which has been linked to mood .

"I like chocolates better than I do cucumbers. They just taste better! I'm just drawn to the taste… you know a banana is lovely, but it doesn't cheer me up like Jaffa Cakes do!" (P8)
Other aspects of food enjoyment were linked to palatability and taste, which were recognised as important in determining food choice, as has been observed across various studies (Yeomans et al., 2004). Visual cues, smells and textures also influenced decisions and determined the level of enjoyment derived from food, which included both savoury and sweet products. A large number of women derived particular satisfaction from sweet products. Sweetness can be a potent stimulus, and research has shown an aversion to bitter tastes and preference for sweetness are embedded in basic biology at birth (Keskitalo et al., 2007). For some women in this study, naturally occurring sugars in fruit and vegetables were not always understood to satisfy cravings. Women distinguished between the potent effects of sugary foods in contrast to fruits and vegetables, which spans biological and psychological levels. The former relates to the effects of sugar as a potent reward that stimulates pleasure responses in the brain, yet may implicate advertising and marketing discourses that present chocolate and sugary snacks as synonymous with relaxation and reward (Locher et al., 2005). A preference for energy-dense foods has long coloured human history, yet the nutritional and physical activity characteristics of hunter-gatherer societies where food functioned as basic nutrition to sustain life has shifted to eating as a socio-cultural construct (Otter, 2012). Thus, while a large number of women enjoyed healthy foods, some were driven by the concept of sugary and fatty foods as treats.

"In my struggles with my emotions and body, common to most women I would think, is body image, and feeling at times loved and at others unlovable" (P4) "…a lot of the time I'll make a choice like not to eat a McDonalds because I'm worried about how it'll affect my body, just how it looks …I feel anxious and tend not to eat…a lot of times if I feel anxious I will just will skip a meal" (P17)
While food was described as an enjoyable part of the women's lives, body image was a powerful aspect of the women's narratives, and eating behaviour was deeply entwined with self-esteem, as correlates with established research (Bedford and Johnson, 2006). Societal pressures, weight fluctuations, and ongoing 'battles' with weight were linked to health concerns. However, ideas of wellbeing were also tied to body image which fuelled maladaptive eating behaviours, such as eating sweets when hungry to provide sugar but fewer calories in full meals, restricting food intake, skipping meals, and exercising to offset overeating. The present sample contained a wide range of ages and body image remained an influential factor in food choice and post-prandial mood at all ages, which agrees with research indicating its significance across a woman's life course (Bedford and Johnson, 2006).

"If I overeat, I feel very upset …later I feel so guilty and so bad… because affects my mood, affects sleep, and it affects everything so then I feel like terrible…and then no one like to be close to me" (P18)
Body image anxieties were often described against the backdrop of the guilt which often followed spells of overeating. Excess calories and fat have been linked to negative mood state in participants for up to two days post-consumption (Hendy, 2012), with results indicative of food as the first influencer in the food-mood relationship. In the present study, guilt followed overconsumption which was associated with lack of control, and suggests food may come first when it relates to guilt. Guilt has also been identified with a propensity to increase or curb food intake following overconsumption of comfort foods and even more nutritious meals (Wansink et al., 2003). However, overeating was generally linked to indulgent foods in this sample, with nutritious meals dissociated from guilt and generally connected with happiness, cooking and togetherness.

Theme 4: Faith and religious beliefs
Intrinsic religiosity was significant in shaping the food-mood relationship within the women's narratives, as deeply rooted beliefs and religious practices intersected with health, nutrition and wellbeing. Faith in a higher Being was connected to ideas of whole person wellness, which included intellectual, physical, mental and emotional elements.

"Food…we know it's a gift from God" (P19)
Ideas of food as 'gifted' from God interconnected with social participation via shared food, which was recognised as a core aspect of church culture that generated joy. Positive perceptions of eating were thus associated with Christian beliefs and practices, with value found in sharing food and 'blessing' food together, which reinforced the women's sense of connection to others and feelings of belonging. This relational aspect of the women's faith thus strengthened their positive life experiences and wellbeing. Previous research links enhanced social support networks to religious involvement (Prado et al., 2004).

"I try not to eat as much meat… it goes back to looking at Genesis (biblical book) and knowing that they weren't killing animals and eating them in the beginning, that wasn't the design. The original first design was just to eat off the land which was the plants and the fruits in the Garden of Eden" (P7)
Heeding biblical messages was deemed important for establishing a healthy approach to diet and lifestyle in the women's daily contexts. Scripture was understood as a signpost for positive lifestyle behaviours that provided an interpretive framework to view diet and nutrition. The belief that employing a Biblical foundation to eating can help take care of one's body was identified as a means of increasing self-efficacy in health-promoting behaviours and as contributing to healthier eating choices. Such findings are consistent with the literature which links religiosity with increased sense of control (Koenig, 2012), as scriptural guidance seemingly afforded the women a sense of regulation and direction in their decisions.

"My journals over the years record a story of me talking to Jesus about every area of my life including my health, weight, living healthily and aiming to honour Him who made me and gave me breath. Jesus has helped me as I've become more secure in him, his love and purposes for my life" (P4)
Participants reflected on their sense of identity, value and purpose in a higher Being who 'loves' them and 'cares' about their wellbeing. This was understood to provide motivation, strength and direction in carrying out and prioritising health supporting behaviours, as well as feelings of security, contentment and solace. The idea that faith fostered identity and purpose added value to themselves and their daily actions, and was described as having a directly positive impact on mood and personal growth. Such narratives are supported in literature that indicates intrinsic religiosity may be significant for improved individual mental health (Amadi et al., 2016).

Limitations and recommendations for research
Unravelling conscious affective mechanisms in the food-mood matrix through perceptions and experiences is both complex and problematic, and a host of variables influenced the data garnered from this study. Part of the complexity of mood state lies not only within its subjective nature, but due to the variety of moods that lie across the affect spectrum. While some patterns could be traced, such as the connection between environments, convenience foods and time, unique factors pertinent to individuals may have amplified connections, such as barriers to accessing supermarkets or working patterns. This ethnically diverse sample of women was drawn from a borough which has both prosperity and socioeconomic deprivation and thus responses within the groups may reflect different opportunities, capabilities and motivations (Michie et al., 2011). Further, the sample contained a wide range of age groups, and as such responses reflect differing experiences across the life course which may inhibit cohesiveness in findings. An additional consideration for future studies includes examining if BMI has any bearing on food-mood responses, due to connections between obesity and mood disorders (Luppino et al., 2010).

Implications
The improvement of mental health and dietary health are public health goals (London Health Commission, 2015). For the UK government to meet its commitments to public mental health, targeting social and environmental components of the food domain may provide opportunity to enhance this relationship, through the creation of relevant services and strategies that foster women's mental health, and in doing so the health of their families and communities. Encouraging more communal eating events may improve health outcomes for the family and the community.

Conclusions
This study explored the complex linkages between food and mood as experienced among women in one borough of London. Questions remain about how to modify relationships with food that improve mental health and wellbeing among women. Further studies are warranted to gain a comprehensive assessment of the relationship between food and mood among women in other settings.