Pregnancy diet based on ancestral patterns increases growth in subcortical fetal brain regions Authors: Lora L. Iannotti, Gabriela Vintimilla Andrade, Shirley Tipanquiza Piedra, Rachel B. Zimmerman, Elizabeth Hahn, Pamela Camana, Dessire Baldeon, Paul Silva, Karen Balladares, Carlos Andres Gallegos-Riofrio, Melissa Chapnick, Gifty Aboagye-Mensah, Timothy Zielonko, Marcus Raichle, Manu S. Goyal, William F. Waters, Iván Palacios León, Jennifer Nicholas Published: 2026 ABSTRACT SignificanceDietary patterns today have diverged substantially from our hominin ancestors, with repercussions for human health, notably brain development. The Mikhuna (“nourish” in Kichwa) trial tested a diet during pregnancy modeled on evidence for dietary patterns in evolution and adapted to the local context of Ecuador for fetal growth in brain and bone outcomes. The intervention, consisting of a variety of sustainably sourced healthy foods (fish, eggs, fruits, and vegetables) and a social behavior change communication strategy, significantly increased maternal diet diversity. Ultrasound imaging captured the Mikhuna effects on fetal brain growth in the corpus callosum and gangliothalamic ovoid with implications for hemisphere connectivity, signal relaying, and other enabling and regulatory functions. Globally, 22.5% of children experience stunting early in life with serious repercussions for survival, health, and neurodevelopment over the life course (1). Stunting is a form of chronic malnutrition arising from poverty, disease, and poor quality diets. A child who is stunted is at increased risk of mortality, infectious and chronic diseases, impaired development across multiple domains, diminished schooling and lifetime earnings, and may even have intergenerational health consequences (2, 3). Nutrition during the first 1,000 d of life, from conception through 2 y of age, is paramount to establishing growth and development trajectories across the life course (4). Brain development particularly unfolds at a rapid pace, necessitating certain nutrients and bioactive factors to catalyze and undergird this growth (5). While the evidence base supporting nutrition and early neurodevelopment has comprehensively examined child functional outcomes, it remains largely uncharted in terms of the nuances of brain biology (6). Here, we use ultrasound imaging to examine maternal nutrition effects on regional brain morphologies and bone growth in the offspring. Dietary patterns globally have diverged substantially from our hominin ancestors, dating back some two million years (7). For most of our time on Earth, Homo sapiens evolved on diets characterized by high dietary diversity and the presence of certain food types, including fresh fruits, vegetables, aquatic and other animal source foods—features also associated with brain size and morphology, at particular junctures in evolution (8–10). The true Paleolithic dietary pattern, distinct from the fad Paleo diet today (11) and reflecting gathering, hunting, and fishing activities, was context and climate dependent. Relative to dietary patterns today, the Paleolithic diet generally included: higher protein and fats, coming from lean game meats, seafoods, shore bird and reptile eggs; lower carbohydrates, coming from tubers, fruits and vegetables, and honey; and higher micronutrients and fiber, coming from a great variety of fruits and vegetables (8). Contemporary diets, by contrast, contain high proportions of carbohydrate-based staples, and in some populations, widespread consumption of ultraprocessed foods (UPF) (12, 13). Child growth and brain development are endangered by modern diets and the related problem of hidden hunger, or nutrient deficiencies in iron, zinc, iodine, vitamin A, vitamin B12, among others (14, 15). These nutrients are vitally important early in life and are often found lacking in poor quality diets. Strong evidence suggests the harmful effects of these trends, though few intervention studies have examined locally adapted healthy dietary patterns beyond the Mediterranean or Dietary Approaches to Stop Hypertension (DASH) diets, unsustainable in many parts of the world (16–19). Ecuador, considered a middle-income country, presents the hallmark signs of nutrition transition: persistent high levels of stunting combined with increasing prevalence of overweight and obesity, and widespread nutrient deficiencies. The double burden of malnutrition in the country was described over a decade ago with the overlapping problems of stunting, overweight, anemia, and/or zinc deficiencies affecting over 13% of households (20). Micronutrients, including iron and zinc, play critical roles in early brain development processes such as myelination and neurogenesis (14, 21, 22). Ecuador is also among the world’s most important biodiversity hotspots, though threatened by development and environmental degradation processes. Drawing on evolutionary theories and empirical evidence from the disciplines of anthropology, archeology, and public health, the Mikhuna (“nourish” in Kichwa, the local Indigenous language) trial applied a dietary pattern using locally available and sustainable foods in Ecuador (7–10, 23). We developed the pattern intervention based on a gap analysis and our extensive previous research over a decade on culturally appropriate and sustainable foods (24, 25). Here, we tested the efficacy of a maternal dietary pattern (fish, eggs, fruits, and vegetables) plus social marketing for behavior change on fetal and newborn growth and brain development. Results Enrollment characteristics were comparable for control and Mikhuna groups (SI Appendix, Table S1). One-quarter of the participants were primiparous, experiencing their first pregnancy. The majority of women reported having completed secondary education, while less than half were working outside the home. Overall, there was a 19.5% loss-to-follow-up over the course of the trial, with higher losses in the control group at both end points assessed, 35 wk and 2 wk postpartum (SI Appendix, Fig. S1). The most common reasons were miscarriage (n = 17), moving from the area (n = 13), and lack of time (n = 6). Mothers lost-to-follow-up were on average approximately 2 y younger, 24.7 ± 6.01 y, than mothers remaining in the trial, but no other differences in enrollment characteristics were detected. The Mikhuna intervention was effectively delivered, consistent with the trial protocols, and showed high adherence to the dietary pattern intervention (SI Appendix, Table S2). Workshop attendance was lower than expected, with 63.5% of Mikhuna participants and 60.3% of control participants attending at least one workshop. Dietary diversity was significantly improved in the Mikhuna intervention group compared to the control, with a greater percentage of women meeting Minimum Dietary Diversity for Women (MDD-W) threshold in Mikhuna (74.5%) vs. control group (55.8%) (P = 0.004) (Fig. 1). The total number of distinct, individual foods consumed in a day was also significantly increased in Mikhuna (14.0 ± 9.4) vs. control (9.0 ± 8.2) (P < 0.001), and the percentages reporting at least one food in each MDD-W food group were significantly higher in Mikhuna vs. control, for 7 of the 10 food groups (Fig. 1). Any meat consumption and frequency of meat consumption did not differ by treatment group and was not associated with trial outcomes. There remained high prevalences of both rice (96.7%) and UPF consumption (92.8%) at 35 wk gestation, with no evident differences by group. Dietary diversity among control and Mikhuna group participants at close of intervention, 35 wk gestation. (A) Percentage of participants reporting any consumption of individual foods within 10 food groups in a 24-h period. Food groups were developed by FAO for the MDD-W indicator. Gray bars indicate foods with no reported consumption. (B) Percentage of participants reporting consumption of at least one food within food groups. * Significant difference between groups by chi squared P < 0.05. Maternal morbidities did not differ by treatment group with the exception of nausea and vomiting prevalence at 21 wk and fever prevalence at 35 wk showing higher percentages in the Mikhuna group compared to control (SI Appendix, Table S3). Mothers gained an average of 8.47 kg (SD 3.49) from 12 wk to 35 wk, with no difference evident by treatment group. Fetal growth in bone and brain parameters was significantly different between groups for corpus callosum, gangliothalamic ovoid (GTO) height, and femur length from 21 wk to 35 wk, after adjustment for gestational age at measure and parameter at 21 wk (Figs. 2 and 3). Standardized measures in Z scores also showed significant differences between Mikhuna and control at 35 wk in corpus callosum 0.56 95% CI ([0.02, 1.10]), P < 0.05, and femur length −0.21 (95% CI [−0.42, 0.00]), P < 0.05 at 35 wk (SI Appendix, Table S4). Newborn anthropometry showed nonsignificant differences between Mikhuna vs. control groups [mean (SD)]: length-for-age Z score (LAZ) [−0.81 (0.88) vs. −0.86 (0.99)], weight-for-age Z score (WAZ) [−0.68 (0.99) vs. −0.76 (0.92)], and weight-for-length Z score (WLZ) [−0.26 (1.05) vs. −0.33 (1.00)] (SI Appendix, Table S5). Growth in fetal bone, brain, and placental structures in Mikhuna compared to control group participants from 21 wk to 35 wk gestation. Values indicate mean (SD) or median (min, max) difference for each study arm. Forest plots show difference-in-difference group coefficient from generalized linear models adjusted for gestational age and same parameter at 21 wk. * Significant differences in parameter, P < 0.05. Growth distributions and case example ultrasound images for brain and bone parameters showing significant differences in Mikhuna vs. control groups by generalized linear regression (GLM) models. (A) Case example ultrasound scan for GTO at 35 wk, distribution and scatter plot for difference in GTO height (cm) in control and Mikhuna groups. (B) Case example ultrasound scan for sagittal view of corpus callosum at 35 wk, distribution and scatter plot for difference in corpus callosum length (cm) in control and Mikhuna groups. (C) Case example ultrasound scan for femur at 35 wk, distribution and scatter plot for difference in femur length (cm) in control and Mikhuna groups. In ancillary analyses, the intervention effects did not differ in the stratified sample of term infants compared to the full sample reported here (preterm plus term). Primiparity also did not mediate or modify the intervention effects, though it was significantly associated with increased growth in GTO area and decreased femur length and Z score. Among mothers without gestational diabetes (n = 135), findings were comparable to the full sample models, though greater effect sizes were evident in the growth of corpus callosum from 21 wk to 35 wk, 0.21 cm (95% CI [0.04, 0.39]) and corpus callosum Z score, 0.65 (95% CI [0.08, 1.20]). One additional subcortical parameter showed significant increases in growth from 21 wk to 35 wk: thalamus length, 0.15 cm (95% CI [0.03, 0.27]). Discussion The Mikhuna trial evaluated the impacts of a maternal dietary pattern—modeled on empirical evidence from nutrition in human evolution and adapted locally—on fetal and newborn growth and development. Ultrasound protocols enabled the investigation of macrostructural parameters deeper in the brain than the standardly captured cortical features. The Mikhuna intervention showed significant increases in subcortical brain regions of corpus callosum and GTO, and decreased femur bone length at 35 wk. While these findings supported our brain morphology hypothesis, we did not find differences in newborn anthropometric outcomes. Here, we uniquely blend the disciplines of public health nutrition, neuroradiology, and agroecology to demonstrate the effects of a sustainable dietary pattern on fetal growth and development, showing impacts on brain regions that play critical roles in hemisphere connectivity, signal relaying, and other regulatory functions (26, 27). Sustainable Healthy Dietary Patterns. We set out to design and test the effects of a holistic dietary pattern to build on the preponderance of evidence in public health nutrition focused on single nutrients or even individual foods. Drawing on evidence from multiple disciplines, we applied an evolutionary lens to develop the dietary pattern adapted to the local Ecuadorian context (8, 10). A salient theme in this literature describes changes in brain size and morphology at junctures coinciding with dietary shifts. The shore-based paradigm highlights the contribution of foods such as fish, mollusks, crustaceans, shore-bird eggs, and aquatic plants for providing brain-selective nutrients and lipids, particularly to meet energetic demands for early brain development (9). Fish was a central feature in the Mikhuna food basket and may have contributed to the brain morphological effects, notably for the white matter corpus callosum. One study showed maternal intakes of omega-3 fatty acids at 1 mo postpartum positively correlated with infant corpus callosum volume (28). Our study did not account for the possible polymorphisms in the genes encoding for fatty acid the desaturase enzyme involved in long-chain polyunsaturated fatty acid (PUFA) production. The biological effects of PUFA may differ by genetic variants (29). Another issue related to fish and human nutrition is the potential for heavy metal contamination, particularly in predator fish high in the food chains (30). Though to our knowledge, the trout, tilapia, and canned tuna used in Mikhuna originated from safe sources (31). Environmental sustainability was a key criterion in designing the Mikhuna dietary pattern. Foods were sourced from agroecological farms with a view toward scaling sustainably and strengthening local food systems. By integrating native fruits (e.g., mora/Rubus glaucus; naranjilla/Solanum quitoense) and vegetables (e.g., zapallo/Cucurbita ecuadorensis; achojcha/Cyclanthera pedata), Mikhuna leveraged Ecuador’s rich biodiversity while fostering resilience (Fig. 4). We assessed changes in food biodiversity using a proxy marker for species richness or the total count of individual foods consumed (32, 33). This marker and the MDD-W showed significant increases in the intervention group. However, some challenges did emerge when unfamiliar foods were introduced into the weekly baskets, requiring the study team to provide mothers with recipes and preparation ideas to encourage their adoption. Furthermore, UPF and heavy rice consumption remained in the diets of nearly all the mothers from both control and Mikhuna intervention groups. Finally, with regard to environmental sustainability, climate change has affected fish availability and aquaculture in the region, particularly as diminishing spring water sources continue to disrupt local ecosystems (34). Mikhuna food biodiversity. The map of Ecuador shows the location of the two study sites in the communities of El Quinche and Yaruquí. Native foods distributed in the Mikhuna food basket are listed by common and scientific names and grouped in regions of origin: Coast (blue); Highlands (brown); and Amazon (green). Mikhuna and the Evidence Base. The literature interrogating nutrition-brain development relationships largely comes from animal models and observational studies focusing on single nutrients, populations from high-income contexts, and later phases of the life course after insults have occurred (6, 35). The Generation R prospective cohort study, carried out in the Netherlands, served as a basis for some of our ultrasound protocols and, importantly, for the hypothesis related to the GTO. This study involving 9,778 mothers was designed to examine genetic and environmental factors related to growth and development from fetal life to young adulthood (36). One analysis embedded within Generation R showed exclusive breastfeeding was associated with larger GTO diameter (difference in means 0.21 cm (95% [CI 0.02, 0.39]), P = 0.02) at 7 wk postnatal (37). Another potentially relevant observational study linked early childhood stunted growth with brain morphometric measures in children at 9 y of age (38). They showed a declining trend among children “not stunted” to “always stunted” for size of similar brain regions significantly impacted by the Mikhuna intervention—the corpus callosum and GTO subregions (right pallidum, bilateral caudate, and right thalamus). The Mikhuna study also corroborated findings from a similar intervention trial conducted in Spain (39). Their study targeted mothers at risk for small-for-gestational age infants and examined the effects of the Mediterranean diet compared to mindfulness-based stress reduction on fetal and neonatal brain development using MRI measures. At 38 wk, adjusted regression models showed increased corpus callosum size, as well as total fetal brain volume and right frontal lobe in the Mediterranean diet compared to the usual care group. Offspring from the mindfulness-based stress reduction group showed a larger fetal left anterior cingulate gyri volume compared to the usual care group (39). Implications of Brain Effects. A critical goal of maternal nutrition interventions is to optimize future neurological outcomes, including cognitive function, mental health, and overall well-being. Intermediate biomarkers in the offspring and endophenotypes can offer early insights into growth and development trajectories and guide interventions on modifiable factors. Corpus callosum biometry has been studied previously as a potential biomarker predictive of neurodevelopmental outcomes (40, 41). As a structure that connects homologous cortical areas of the left and right cerebral hemispheres, the corpus callosum helps to orchestrate a complex set of sensory, motor, and cognitive functions (26). It is the largest white matter tract in the human brain, thereby reflecting in part white matter viability and functional networks more broadly (42). Nutrition effects on white matter development has been largely studied in preterm infants (43), with evidence indicating positive impacts from early lipid intakes (6). These findings suggest the fish, rich in long-chain fatty acids, provided in the Mikhuna intervention may have played a role in the increased size of corpus callosum. The GTO comprises both basal ganglia and thalamic structures, which are involved in multiple brain networks, including those relevant to motor control, sensory gating, and cognition. Among the most highly connected gray matter structures in the human brain, the basal ganglia and thalami also develop early in life, being highly metabolically active during the initial months of life. Evidence is more limited on biometric measures of the GTO and their relevance to neurodevelopment. The Generation R study demonstrated that a smaller GTO was associated with increased internalizing problems and subscales of anxiety and affective problems (44). In other analyses, they found a larger GTO in infants aged 6 wk was associated with a lower risk of attachment disorganization at 14 mo (27) and that exclusive breastfeeding was associated with increased GTO size (37). Future investigations are needed to determine how the Mikhuna dietary pattern influences neurodevelopmental outcomes through childhood and beyond. Implications of Bone Effects. Growth in the femur was significantly reduced in the Mikhuna group compared to the control at 35 wk. Although the effect did not translate into differences in LAZ or WLZ at birth, we explored reasons for the bone growth difference. One potential explanation could have been food displacement arising from the Mikhuna dietary pattern. In other words, mothers could have decreased or completely eliminated foods that promote bone growth when adding the intervention foods into their diets. For example, maternal milk and dairy consumption has been found to be associated with fetal femur length and birth weight and length (45). However, maternal milk and dairy product frequency of intakes did not differ in our sample, nor were there differences in calcium supplement or all supplement reporting. These variables were also tested in relation to femur growth without any associations detected. Another potential explanation could have been reduced energy or micronutrient intakes associated with increased nausea and vomiting in the Mikhuna group relative to the control around 21 wk timepoint (SI Appendix, Table S3). This trial followed the maternal-child dyad for a relatively short period of time and did not allow us to monitor growth plate dynamics and bone development over time. One prospective cohort study in Ethiopia examined fetal measures of femur length, abdominal circumference, biparietal diameter, and head circumference at 26 wk, 30 wk, and 36 wk gestation in relation to anthropometric Z scores at 11 to 24 mo of age (46). They found femur length at 24 wk to 28 wk to be significantly associated with child LAZ at 11 to 24 mo. In this respect, a follow-up cohort study in these children could provide deeper insights into the longer-term bone and brain effects. Ultrasound Imaging for Public Health Research. Ultrasound is a portable, relatively low-cost imaging modality that allowed us to evaluate brain structures not included in a typical prenatal ultrasound at community health centers. Several of the same brain structures were evaluated postnatally, captured by imaging through the fontanel. The utility of portable ultrasound was also demonstrated in our earlier studies carried out in the rural region of Cotopaxi, Ecuador (47, 48). The study team visited households with a portable ultrasound to evaluate the bone age of young children, otherwise logically challenging to execute if families were asked to travel to local clinics for the X-ray. When coupled with the other facets of the Lulun II study, a relationship between stunting and bone age was demonstrated (47). In the Mikhuna trial, we added several brain parameters to those standardly measured during pregnancy, allowing us to explore a greater breadth of macrostructural changes in relation to nutrition. The range in the reliability of measures, based on Cronbach’s alpha, suggests additional training and supervision may be needed to ensure precise and accurate measures of the smaller, deep brain regions in particular. In this context, as in others around the world, such evaluations with sophisticated technologies such as MRI would have been both invasive and prohibitively expensive. Ultimately, the Mikhuna randomized controlled trial (RCT) findings applying difference-in-difference analyses point to internal validity, and parameter distributions were consistent with World Health Organization (WHO) fetal growth curves and previous studies looking at additional features (49–51). Limitations. Some limitations were present for the Mikhuna trial. The losses to follow-up (19.5%) potentially lowered the power to detect some group differences. Additional subgroup analyses (e.g., primiparous vs. multiparous and premature vs. term infants) may have been possible with fewer losses-to-follow-up or a larger sample size. Another limitation in our view was not applying ultrasound estimates of gestational age at the 12 wk enrollment period. Instead, we used the date of the last menstrual period (LMP) and clinical criteria by obstetrician/gynecologist (OB/GYN), and later adjusted gestational age based on ultrasound imaging at 21 wk using the Hadlock equations described previously when outside acceptable ranges. All regression models in this study adjusted for the same ultrasound parameter at 21 wk as a proxy for baseline status. However, this likely attenuated intervention effects since the intervention started at enrollment between 12 wk and 14 wk, when assessing these parameters would not have been possible. Food preferences that arise during pregnancy may have interfered with full adherence to the dietary pattern, especially regarding fish. Other factors related to diet were the high levels of rice and UPF consumption in the sample. Rice, a starchy staple, is integral to the food culture of Ecuador, while UPF are pervasive in the food environment. Finally, we had lower workshop attendance than expected arising from time-scarcity and logistical challenges for the mothers in both groups. Other elements of the social behavior change communication (SBCC) including messaging during the weekly food deliveries and recipe sharing likely supported the increased dietary diversity outcome especially. Conclusions The Mikhuna trial offers evidence on the positive health effects of a diverse and sustainable dietary pattern during pregnancy on fetal development and maternal nutrition. In our view, ongoing research is needed to understand how to adapt dietary patterns to local contexts with a view toward protecting both human and ecosystem health. This study blended expertise across multiple disciplines to design an intervention and probe for biological impacts using noninvasive ultrasound methods. Our team is disseminating Mikhuna findings to various stakeholders in Ecuador and beyond, with the potential for future translation into policy and practice. Methods The Mikhuna trial was an individually randomized controlled trial conducted in two peri-urban, low-resource municipalities of Ecuador, El Quinche, and Yaruquí (Fig. 4). The study protocol was approved by the Human Research Protection Office at Washington University in St. Louis (WashU) (IRB Biomedical ID# 202001038) and at Universidad San Francisco de Quito (USFQ) Ethics Committee for Human Research (ID# 2019-138 M). Informed consent was obtained from all participants in Spanish. The communities were selected based on low socioeconomic status and food insecurity, existing relationships with USFQ, and proximity to Quito. Pregnant women were eligible to participate if they were healthy, at least 18 y of age, carrying a singleton pregnancy, and in their first trimester up to 14 wk of gestation. Gestational age at recruitment was estimated using the LMP date. For some women, gestational age was later adjusted based on clinical criteria by the OB/GYN physician or ultrasound imaging in the first and second trimesters (n = 15), and Hadlock equations (n = 21) (52). Exclusion criteria were the following: less than 18 y; chronic health condition (HIV, tuberculosis, diabetes, cardiovascular conditions, etc.); multiple pregnancy (twins, triplets, etc.); intent to move away from the study area; known allergies to eggs or fish; and blood pressure greater than 160/100. To find and recruit eligible women, the study team used the clinics’ databases and relied on referrals from physicians, nurses, and other healthcare workers. Once eligibility was established, informed consent processes were initiated. Randomization. Those agreeing to participate were randomized in blocks of 10 during the rolling recruitment phase from April 2022 to July 2023, implemented by the field study team. Women selected chips from an envelope to receive their group assignment to either: 1) control group standard-of-care; or 2) Mikhuna intervention. After all 10 chips were selected and women assigned, the chips were returned to the envelope for the next block of women. Sample size (n = 215) was based on variability observed in sonographic measures taken in Ecuador in our previous pilot study (48), and an estimated effect size of 0.15 LAZ from previous maternal nutrition studies. Intervention. The intervention from 12 wk gestation to birth consisted of a weekly food delivery [8 eggs (4 servings); 500 g fish (2 servings); a variety of fruits, vegetables, herbs, and spices] and a SBCC strategy encouraging diet diversity and limiting UPF. Mothers were initially monitored for any potential food allergies, though none were identified. The dietary pattern was designed drawing on evolutionary theories, population food and nutrient gap analyses, and formative research for culturally acceptable foods. Common characteristics of diets in the longest period of hominin evolution (gatherer-hunter-fisher diets) were invoked: diet diversity; specific animal foods (fish, lean meat, eggs); specific plant foods (berries, nuts, tubers, etc.); and no highly processed foods. We developed a food list meeting these characteristics, combined with the criteria of culturally appropriate, sustainably sourced, and filling nutrient (e.g., vitamin A, zinc) and food gaps (e.g., fresh fruits and vegetables) in the current diet. Amazonic Gardens, a local food supplier with sustainable food production practices, delivered the food bags weekly to clinics in El Quinche and Yaruquí. Mothers from the Mikhuna group came to the clinics at designated times to avoid overlap with control group visits. Some minor adjustments were made initially to the food basket—specifically reducing the quantity of fruits and vegetables—based on mothers’ feedback of too much food and rising prices due to strikes, gas prices, and avian influenza. Exact quantities of each delivery were weighed and recorded (SI Appendix, Table S2A). Fish type was determined based on availability and preference; 50.6% of deliveries contained trout, 42.3% tilapia, 2.2% paiche, 3.5% tuna, and 1.2% sardines. The study team distributed the food bags and administered a short survey to assess compliance. To reduce the likelihood of any interactions with control group participants and risk of contamination, a number of precautions were taken. First, weekly food deliveries were scheduled on days when only mothers from the Mikhuna group attended the clinic. Second, plain bags commonly seen in the local market without project identifiers were used for the food basket and, finally, the bags containing foods were kept out of sight (e.g., deliveries were made to a back entrance in Yaruquí and bags were stored in a tent in El Quinche). To design the social marketing element, we conducted a market study on participant norms, values, and symbols, as well as the perception of the brand, including its most visible elements such as the name and visual representation (i.e., logo), together with our previous studies (53). The brand serves as a symbolic representation of the project, acquiring meaning in the minds of those who engage with it. For example, the name Mikhuna—which means “food” or “to nourish” in Kichwa—was chosen to reflect a broader understanding of nourishment. In the Andean Indigenous view, food is not merely sustenance for the body; it is spiritual, for nourishing the soul, and preserving traditions. Integral to the social marketing plan was a set of voluntary workshops designed in large part to encourage participation among all participants and communicate targeted nutrition messages to the Mikhuna group: 1) introduction; 2) emotions in pregnancy; 3) self-esteem in pregnancy; 4) self-care in pregnancy; 5) breastfeeding, gestational diabetes, and workout routine; and 6) baby shower. Mothers were separated by group assignment and progressed through the series with approximately 10 to 12 other mothers. A clinical dietitian conveyed messages to the Mikhuna groups regarding dietary diversity, elimination of UPF, and healthy dietary patterns. Finally, the study team offered guidance to Mikhuna participants periodically on preparation and recipes, mostly by sharing YouTube videos. Adherence to the intervention was tracked through workshop attendance records and a survey module administered weekly at the time of food delivery, asking a series of questions about consumption of the foods in the previous week (SI Appendix, Table S2 A and B). Data Collection. Data for the Mikhuna trial were collected longitudinally using REDCap 15.0.10. All data was double- and triple-checked to ensure accuracy. Data were downloaded from REDCap using the deidentification feature during data export, allowing for the removal of sensitive information and direct identifiers. At baseline 12 wk gestation, we collected socioeconomic and demographic data, maternal morbidities, anthropometry and dietary intakes, and blood. Early in the study, some minor changes were made to the survey questions for the following reasons: to improve the Spanish translation, provide specificity on seafood consumption, improve formatting, and add “other” or “none” options where needed. At 21 wk and 35 wk gestation, maternal morbidities, anthropometry, and dietary intakes were reassessed, while ultrasound protocols were followed. A second blood collection occurred at 35 wk. At 2 wk postpartum, neonatal anthropometry, ultrasound, and neurodevelopmental outcomes were assessed. Anthropometry, Dietary Intakes, and Health. Anthropometric measures of height, weight, and head circumference of the mother and newborn were collected at 2 wk postpartum according to international protocols (54). Pairs of enumerators took two measures of maternal height and newborn length using Infant/Child/Adult Shorrboard ® (Weigh and Measure, LLC, Maryland) to the nearest 1 mm. When measures differed by 5 mm or more, a third measure was taken and averaged with the closest measure. Weight was measured using an electronic digital scale (seca GmbH and Co. K.G., Hamburg, Germany) to the nearest 0.01 kg. Again, two measures were taken, and when differing by 0.1 kg or more, a third measurement was taken and averaged with the closest other measure. Maternal blood pressure was taken using Microlife BP A2 Classic blood pressure monitor, and morbidities were assessed through 2-wk recalls. Maternal morbidities were examined by inquiring whether the participant had symptoms or diagnoses since their last antenatal visit. For maternal dietary intakes, we applied a 24-h food frequency recall method drawing on a comprehensive list of foods from Ecuador developed by USFQ. In addition to the food items, participants were asked if they took any supplements (iron, folic acid, zinc, multi-micronutrient) in the previous 24 h and whether they took any supplements at any timepoint in pregnancy. We calculated the MDD-W based on any consumption from 10 standardized food groups: grain, roots, and tubers; pulses; nuts and seeds; dairy; meat, poultry, and fish; eggs; dark leafy greens and vegetables; other vitamin A-rich fruits and vegetables; other vegetables; and other fruits (55). MDD-W was calculated with the following: formula:Women 15 to 49 y of age who consumed foods from 5 food groups during the previous dayTotal number of women 15 to 49 y of age surveyed. Food biodiversity has been described to include three components: richness, evenness, and disparity (33). Species richness has been shown to be predictive of positive health and nutrition outcomes across different contexts (32). While we did not have species identified for all foods included in the FFQ, the detailed list provided the opportunity to assess total number of distinct foods consumed and compare across groups. Brain and Bone Imaging. The Mindray M7 Portable Ultrasound Machine was used to evaluate fetal and newborn growth and brain development. The co-Principal Investigator, a pediatric radiologist trained the field-based radiologists and residents at USFQ over a 1-wk period on standard measures (biparietal diameter, head circumference, cerebellar diameter, corpus callosum sagittal length, abdominal circumference, humerus and femur length, placenta volume and grade), and new protocols for additional regions [thalamus length, width, area, and transverse; GTO length, height, volume, and area]. Imagining of these subcortical brain features were included to expand the breadth of regions studied in relation to nutrition and based on findings from the Generation R study (25). Size of the corpus callosum has been shown to be predictive of neurodevelopmental outcomes (40), while a small GTO was associated with internalizing problems (anxiety and affective problems) and attachment disorganization (27, 44). During the trial period, the lead radiologist visited the clinic four times to oversee and ensure high-quality imaging. Images, first recorded on the local drive of the ultrasound machine and a password-protected external hard drive, were routinely shared with radiologists for ongoing supervision and data quality monitoring. Fetal anomalies or other adverse conditions identified through ultrasound scanning were referred to as the appropriate medical facilities for care. Ultimately, the scans were analyzed for intervention effect by the radiologists blinded to group assignment for each individual. Ultrasound imaging more reliably captured measures of the larger brain and bone regions across different timepoints compared to the smaller, deeper regions, with Cronbach’s alpha ranges: 0.48 to 0.61 for biparietal diameter, head circumference, arm circumference, humerus, and femur; and 0.07 to 0.24 for cerebellum, corpus callosum, thalamus, and GTO diameter. Statistical Analysis. Standardized WHO fetal growth curves were applied to generate Z scores and percentiles for these ultrasound measures: biparietal diameter; head circumference; abdominal circumference; humerus length; femur length; and estimated fetal weight (51). Z scores and percentiles for additional parameters were generated with references in the literature: corpus callosum (50). Cerebellum (49), and ulna and tibia (56). WHO Growth Standards (54) were applied to generate anthropometry Z scores and percentiles at birth: LAZ; WAZ; and WLZ. Sample size (n = 215) was predetermined by grant budget, timeline, and attainable numbers for recruitment of pregnant women in their first trimester using the monthly average from clinic records. Power calculations were carried out based on the hypothesis for increased GTO in the intervention group and size estimates from a previous study (48). With a mean GTO diameter of 1.85 ± 0.31, hypothesized effect size of 0.15, and α = 0.05, power was estimated at 1-β = 0.93. Intention-to-treat analyses were applied such that data from all participants were examined for group effects regardless of whether they were in full compliance with the protocol. Descriptive and univariate statistics first ascertained whether randomization procedures were effective, and balance was achieved across the study groups. Difference-in-difference analyses compared change in the fetal and newborn parameters across the study groups for brain and bone parameters collected at 21 wk, 35 wk, and 2 wk postpartum. Similarly, we evaluated the difference by group between change in maternal blood biomarkers from baseline to 35 wk gestation. Generalized linear regression (GLM) models with mixed effects were used to allow for nonnormal distribution of parameters. The first set of models compared groups toward the end of the intervention, assessing change in blood, brain, bone, and placental parameters from 21 wk to 35 wk and inclusive of raw measures and reference Z scores. The second set of models examined neonatal outcomes, comparing anthropometric measures and changes from 21 wk gestation to 2 wk postpartum in brain parameters imaged through the fontanel. Models were adjusted for gestational age at the time of the measure and baseline characteristics with strong prognostic characteristics on outcomes, to improve the precision and reliability of treatment effects (57). Although meat was not delivered in the intervention food basket, we adjusted for any consumption and frequency of meat (e.g., empanadas, soups, ground beef, etc.) in models to account for bioavailable iron supply in the diet, as it is a nutrient important in fetal brain development. These covariates were nonsignificant across all models. Similarly, report of iron supplementation at 21 wk and 35 wk did not impact treatment effects. Ancillary analyses were run to examine effects stratified by premature and term births, primiparous and multiparous pregnancies, and gestational diabetes in view of the literature linking these conditions to fetal growth and development (58, 59). The Benjamini–Hochberg procedure was applied to adjust for multiple comparisons (60). Only the GTO effect, difference-in-difference from 21 wk to 35 wk, remained significant. Here, we report the original models to point to trends and serve as a basis for future hypothesis testing. Diagnostics assessed normality of residuals, homoscedasticity, collinearity, and goodness of fit. Akaike and Bayesian Information Criterion compared models and assessed complexity in GLM models. Data analyses were performed with STATA software (version 16.0; StataCorp, College Station, TX). The trial is registered at ClinicalTrials.gov (NCT05242224).Clinical Trial Registry number and website where it was obtainedNCT05242224Registered 2022-02-06https://clinicaltrials.gov/study/NCT05242224?locStr=Ecuador&country=Ecuador&aggFilters=status:act&term=brain&rank=1 Clinical Trial Registry number and website where it was obtained Clinical Trial Registry number and website where it was obtained NCT05242224 NCT05242224 Registered 2022-02-06 Registered 2022-02-06 https://clinicaltrials.gov/study/NCT05242224?locStr=Ecuador&country=Ecuador&aggFilters=status:act&term=brain&rank=1 https://clinicaltrials.gov/study/NCT05242224?locStr=Ecuador&country=Ecuador&aggFilters=status:act&term=brain&rank=1 Supplementary Material Appendix 01 (PDF)