Machine learning for prediction of schizophrenia based on identifying the primary and interaction effects of minor physical anomalies

In the neurodevelopmental model of schizophrenia, minor physical anomalies (MPAs) are considered neurodevelopmental markers of schizophrenia. To date, there has been no research to evaluate the interaction between MPAs. Our study built and used a machine learning model to predict the risk of schizophrenia based on measurements of MPA items and to investigate the potential primary and interaction effects of MPAs. The study included 470 patients with schizophrenia and 354 healthy controls. The models used are classical statistical model, Logistic Regression (LR), and machine leaning models, Decision Tree (DT) and Random Forest (RF). We also plotted two-dimensional scatter diagrams and three-dimensional linear/quadratic discriminant analysis (LDA/QDA) graphs for comparison with the DT dendritic structure. We found that RF had the highest predictive power for schizophrenia (Full-training AUC = 0.97 and 5-fold cross-validation AUC = 0.75). We identified several primary MPAs, such as the mouth region, high palate, furrowed tongue, skull height and mouth width. Quantitative MPA analysis indicated that the higher skull height and the narrower mouth width, the higher the risk of schizophrenia. In the interaction, we further identified that skull height and mouth width, furrowed tongue and skull height, high palate and skull height, and high palate and furrowed tongue, showed significant two-item interactions with schizophrenia. A weak three-item interaction was found between high palate, skull height, and mouth width. In conclusion, we found that the two machine learning methods showed good predictive ability in assessing the risk of schizophrenia using the primary and interaction effects of MPAs.

Minor physical anomalies including palatal rugae pattern and palatal dimensions in children with sickle cell disease: A cross-sectional analytical study

Background

Sickle cell disease (SCD) is the most common hereditary hemoglobinopathy, which delays growth leading to an altered skeleton and craniofacial pattern. Palatal rugae patterning has been considered the regulator of the development of the palate. The purpose of the research work was to study the morphology of the palate, rugae pattern, and its dimensions in SCD children and compare them with healthy normal children, and to evaluate its role as minor physical anomalies (MPAs).

Methods

A cross-sectional case-control study was designed as per STROBE guidelines. The sample comprised 50 children diagnosed with sickle cell disease (Group SCD) and 50 normal healthy children as control (Group C) belonging to the same age group (10-18 years). Dental impressions were made, followed by the pouring of dental casts. The length of the palatal rugae was measured and categorized into primary (>5 mm), secondary (3 mm-5 mm), and fragmentary rugae (<3 mm). The shape of each primary palatal rugae was identified and categorized as curved, wavy, straight, circular and non-specific. Linear and angular measurements of the palatal rugae patterns and palatal dimensions (width, height, area) were measured and recorded.

Results

The total number of palatal rugae and fragmentary rugae was lesser in Group SCD than in Group C (p < 0.05). The depth of the palate was significantly increased, whereas the area of the palate significantly decreased in Group SCD.

Conclusions

The children with SCD showed distinctive palatal rugae patterns and dimensions when compared with normal healthy children that can be attributed as potential MPAs for sickle cell disease. Children with SCD had an under-developed palatal rugae pattern with a deep, narrow and small palate when compared to healthy children.The dimensions of the palatal rugae pattern in SCD showed reduced distance between the incisive papilla and the first and last rugae, indicating a further decrease in the anteroposterior dimensions of the palate. These findings may aid in the early diagnosis and prevention of malocclusion in children with SCD by appropriate interceptive orthodontic treatment.

Minor physical anomalies in bipolar disorder in comparison to healthy controls and schizophrenia: A systematic review and meta-analysis

Minor physical anomalies (MPAs) are markers of abnormalities in early foetal development and are well established findings in schizophrenia. It has been suggested that neurodevelopmental abnormalities might play a role not only in schizophrenia but also in bipolar disorder (BD). Therefore, according to neurodevelopmental theory of BD, one might expect increased prevalence of MPAs in BD. A meta-analysis of 11 studies was conducted to quantitatively review MPAs in BD in comparison to schizophrenia and healthy controls. The current meta-analysis compared MPA scores of 584 BD patients and 723 healthy controls, and 401 BD and 612 schizophrenia patients. Patients with BD had significantly higher MPA scores than healthy controls (g=0.47, CI=0.28-0.67). This was true both for craniofacial (g=0.57, CI=0.34-0.79) and periphery (g=0.46, CI=0.18-0.73) MPAs. BD was associated with a less severe increase in MPA score compared to schizophrenia, however, between-group difference was small (g=0.19, CI=0.05-0.33). The outcome of this meta-analysis suggests that BD is associated with medium effect size increase in MPAs which is only minimally less severe than schizophrenia. This finding supports the hypothesis that early developmental insult in brain development plays a role not only in schizophrenia but also BD. Studies investigating clinical, neurocognitive, neuroanatomical and other biological correlates of MPAs in BD might helpful in characterizing subtype (s) of BD that is associated with pronounced deviations in brain development.

Cross-diagnostic evaluation of minor physical anomalies in psychiatric disorders

BACKGROUND

Minor physical anomalies (MPA) are markers of impaired neurodevelopment during the prenatal stage. Assessing MPA across psychiatric disorders may help understand their shared nature. In addition, MPA in family members would indicate a shared liability and endophenotype potential. We examined familial aggregation of MPA and their role as transdiagnostic and disorder-specific markers of 5 major psychiatric/neuropsychiatric conditions (schizophrenia, bipolar disorder, substance dependence, obsessive-compulsive disorder, and Alzheimer's dementia).

METHODS

Modified Waldrop's MPA scale was applied on 1321 individuals from 439 transdiagnostic multiplex families and 125 healthy population controls (HC). Stage of fetal development (morphogenetic/phenogenetic)- and anatomical location (craniofacial/peripheral)-based sub-scores were calculated. Familiality and endophenotypic potential of MPA were analyzed with serial negative binomial mixed-effect regression. Cross-diagnostic differences and the effect of family history density (FHD) of each diagnosis on MPA were assessed. Mixed-effects Cox models estimated the influence of MPA on age-at-onset of illness (AAO).

RESULTS

MPA were found to be heritable in families with psychiatric disorders, with a familiality of 0.52. MPA were higher in psychotic disorders after controlling for effects of sex and intrafamilial correlation. Morphogenetic variant MPA was noted to be lower in dementia in comparison to HC. FHD of schizophrenia and bipolar disorder predicted higher, and that of dementia and substance dependence predicted lower MPA. MPA brought forward the AAO [HR:1.07 (1.03-1.11)], and this was more apparent in psychotic disorders.

CONCLUSION

MPA are transmissible in families, are specifically related to the risk of developing psychoses, and predict an earlier age at onset. Neurodevelopmentally informed classification of MPA has the potential to enhance the etiopathogenic and translational understanding of psychiatric disorders.

Clinical utility of biomarkers of the hand in the diagnosis of schizophrenia

A number of biomarkers were assessed in photos and prints of the hands of 95 patients with a variety of mental disorders to determine whether patients with schizophrenia could be distinguished from the others. Patients were recruited as consecutive admissions from an outpatient psychiatric day hospital population. Fourteen patients were diagnosed with schizophrenia/schizoaffective disorder and 81 were diagnosed with other mental disorders. A discriminant analysis yielded an overall 80% correct classification, with a sensitivity (schizophrenia patients identified correctly) of 78.6% and a specificity (non-schizophrenia patients identified correctly) of 80.2%. Significant differences were noted in the proximal interphalangeal joint, eponychium of the middle digit and fingernails. To determine biomarker frequency distribution patients with bipolar disorder were then compared to those with schizophrenia/schizoaffective disorder and then to patients with PTSD. The former yielded an overall 78.6% correct classification, with a sensitivity of 71.4% and a specificity of 85.7% and with similar biomarker frequency distribution for bipolar disorder as for the entire non schizophrenia group. The latter comparison yielded an overall 58.6% correct classification, with no significant differences between the features. The application of these biomarkers in clinical practice could constitute an additional tool for the psychiatrist in cases lacking diagnostic clarity.

Minor physical anomalies in neurodevelopmental disorders: a twin study

BACKGROUND

Minor physical anomalies (MPAs) are subtle anatomical deviations in one's appearance and may suggest altered embryogenesis. MPAs have been shown to be more common in neurodevelopmental disorders (NDDs) compared with typical development. Still, further studies are needed on MPAs in NDDs, especially using twins to adjust for confounding familial factors.

METHODS

Clinical assessments were conducted on 116 twins (61 NDD, 55 controls) from 51 monozygotic and 7 dizygotic pairs to examine MPAs and their association with DSM-5 defined NDDs. Additionally, the relationship between the number of MPAs within twins by zygosity was investigated.

RESULTS

Within the cohort sample, a specific association was found between MPAs and autism spectrum disorder (ASD) diagnosis (crude odds ratio = 1.29, p = .047; adjusted odds ratios = 1.26-1.33, adjusted p values = .032-.073) and autistic traits (crude β = 3.02, p = .002; adjusted β = 2.28, p = .019), but not NDDs in general or ADHD, nor within-pairs. Identified MPAs in ASD included overweight, hypermobility, pes planus, straight eyebrows, vision impairment, arachnodactyly/long toes, long eyelashes, and microtia. The number of MPAs within all monozygotic pairs was highly correlated (r = .88, p < .001).

CONCLUSION

MPAs are more frequent in participants with ASD and may be influenced by genetics. The value of MPAs for (early) detection should be further explored, as they might index individuals at increased risk for ASD in particular.

Iris structure and minor physical anomalies in schizophrenia

This study compared five human iris characteristics and minor physical anomalies (MPAs) between patients with schizophrenia (n = 32) and controls (n = 31). Correlations between iris characteristics and MPAs were expected, due to their same ectodermic origin. Iris macro photos were taken and quantified in five categories mentioned before. MPAs were also examined in both groups. Our results show significant differences in the frequency of pigment dots of the iris and total number of MPAs between groups. Other significant differences were found in the extension of concentric furrows, as they were more common in healthy subjects, while Wolfflin nodules occurred significantly more often in patients with schizophrenia. Expected difference in Fuch's crypts could not be observed between groups. Light eye color was positively correlated to pigment dots and Wolfflin nodules, and negatively correlated with concentric furrows. Dark eye color showed positive correlation with concentric furrows, and negative correlation with pigment dots and concentric furrows. A gender effect could also been observed: male individuals showed moderate positive correlations between pigment dots and total MPAs frequency, while this couldn't be observed in the female group. Our findings suggest possible connections between iris characteristics and MPAs, where males are more prone to deviations.

Palate size and shape in schizophrenia

The palate is considered typical of the structures in which schizophrenia-related minor physical anomalies may occur. In this study, we aimed to compare the dimensions and form of palate in patients with schizophrenia with nonpsychiatric controls in a blinded manner. Dental stone casts of 127 patients with schizophrenia and 127 controls were prepared from impressions of the maxillary dental arch. Palate dimensions were measured on the stone casts using a digital caliper and palatometer. Palate length did not differ significantly between the groups, but there was a significant difference in palate width and depth, which were significantly higher in the schizophrenia group. As a result of using multivariate analysis for assessing independent risk factors affecting patients with schizophrenia, furrowed palate shape, palate width, and ellipsoid maxillary dental arch shape were found to be significant. This study also revealed that patients with schizophrenia demonstrate certain gender-related predilections in the differences of palate parameters compared to same-sex controls. As the palate develops in conjunction with both the face and brain, our study findings can significantly contribute to the assumption that there might be structural abnormalities of the palate that could represent specific markers of embryological dysmorphogenesis underlying schizophrenia.

Minor Physical Anomalies as a Window into the Prenatal Origins of Pedophilia

Evidence is steadily accumulating to support a neurodevelopmental basis for pedophilia. This includes increased incidence of non-right-handedness, which is a result primarily of prenatal neural development and solidified very early in life. Minor physical anomalies (MPAs; superficial deviations from typical morphological development, such as un-detached earlobes) also develop only prenatally, suggesting them as another potential marker of atypical physiological development during the prenatal period among pedophiles. This study administered the Waldrop Physical Anomaly Scale to assess the prevalence of MPAs in a clinical sample of men referred for assessment following a sexual assault, or another illegal or clinically significant sexual behavior. Significant associations emerged between MPA indices and indicators of pedophilia, including penile responses to depictions of children, number of child victims, and possession of child pornography. Moreover, greater sexual attraction to children was associated with an elevated craniofacial-to-peripheral anomalies ratio. The overall sample demonstrated a greater number of MPAs relative to prior samples of individuals with schizophrenia as well as to healthy controls.

Do biometric parameters of the hand differentiate schizophrenia from other psychiatric disorders? A comparative evaluation using three mental health modules

The link between schizophrenia and anomalies in the distal upper limb is well documented. Preliminary studies have identified a number of biometric parameters of the hand by which schizophrenics can be distinguished from matched controls. The current study seeks to determine whether patients with schizophrenia can be singled out from a disparate group of other mental disorders by using the same parameters. We studied three groups, totaling 134 men: 51 diagnosed with schizophrenia, 29 with anxiety and mood disorders, and 54 comprising a control group. Seven parameters were studied: the proximal interphalangeal joint, the eponychia of the middle and ring digits, two dermatoglyphic features, and two constitutional factors. Examiners evaluated the parameters based on photographs and prints. An initial Mann Whitney comparison showed no significant difference between the control group and those identified with anxiety and mood disorders. We therefore accounted for them as a single group. In a discriminant analysis, an overall accuracy of 78.4% was established with a sensitivity of 80.4% (schizophrenics identified correctly) and a specificity of 77.1% (controls identified correctly). These results suggest that the biometric parameters employed may be useful in identifying patients with schizophrenia from a disparate group of other mental disorders.

Evaluation of spontaneous dense array gamma oscillatory activity and minor physical anomalies as a composite neurodevelopmental endophenotype in schizophrenia

OBJECTIVE

Minor physical anomalies (MPAs) and gamma oscillatory activity have been proposed as associated endophenotypes in schizophrenia. Combining these endophenotypes to create a composite endophenotype may help identify those at risk for schizophrenia better. The present study aims to investigate MPAs and gamma oscillatory activity in schizophrenia patients, their unaffected first degree relatives and healthy controls and appreciate whether they can be used together as a composite endophenotype.

METHODS

This was a cross sectional family study conducted at a tertiary care mental health setup. Ninety participants including schizophrenia patients, their first degree relatives and controls (thirty each) were assessed for MPAs on the Extended Waldrop Scale. All participants underwent an awake, resting 192-channel EEG recording. Spectral power and coherence in 30-100Hz gamma bands were estimated using Welch's averaged periodogram method. One-way ANOVA, chi square test were used for comparing socio-demographic-clinical variables. MANOVA supplemented by one-way ANOVAs (post hoc Tukey HSD) were done for comparison of spectral measures. Pearson's correlation, step-by-step linear discriminant functional and intra-familial correlation analysis were subsequently performed.

RESULTS

An endophenotype pattern of finding was found for MPAs in the craniofacial region, the total number of MPAs, spectral power in right temporal region on all bands and in the right parietal region in 50-70Hz and 70-100Hz gamma bands. The three groups were most accurately classified when MPA total score, right temporal 30-50Hz gamma power and right occipital 'intra hemispheric' 50-70Hz gamma coherence were considered together than when considered independently. Significant intra familial correlation was seen for MPA total score and right temporal gamma 30-50Hz power.

CONCLUSION

Composite evaluation of two developmentally linked markers i.e. MPAs and gamma spectral measures may prove useful in categorizing schizophrenia and identifying at-risk individuals.

A Pilot Study: Association between Minor Physical Anomalies in Childhood and Future Mental Problems

OBJECTIVE

This study aims to investigate association between early recognizable minor physical abnormality (MPA) during childhood is associated with mental health problems in young adults.

METHODS

In 1984, 169 preschool children in central Taiwan underwent a detailed physical examination for subtle abnormalities (MPA). Fourteen years later, the Brief Symptom Rating Scale (BSRS) and Chinese Health Questionnaire (CHQ) were used to measure specific psychiatric symptoms.

RESULTS

There is an association between MPA during childhood and adult characterized with interpersonal sensitivity, anxiety, depression and paranoid mental health symptoms.

CONCLUSION

The signs of childhood MPA can be easily identified and should be regarded as risk factors when predicting mental disorder. Mental health professionals should consider MPAs as important signs for possible development of emotional problems.

Biometric parameters of the hand as an index of schizophrenia--a preliminary study

Since abnormalities in distal upper limb development are among the minor physical anomalies associated with schizophrenia we attempted to determine whether patients with schizophrenia can be identified on the basis of specific morphologic and dermatoglyphic features of the hand. Photographs and prints of the hands of 38 patients with schizophrenia and those of 42 control subjects were evaluated and graded on 13 biometric parameters. Results were statistically evaluated. A combination of three of the parameters was found to have good predicting abilities to distinguish between schizophrenics and controls. Subjects having high values in these three parameters were found to have a higher propensity to be defined as schizophrenics. In order to define a simple rule for classifying subjects we chose a criterion of having a value of 3 (in a scale from 1 to 3) in at least one of these three discriminating variables. This rule yielded an overall accuracy of 81.2%. Among controls, 85.7% of subjects did not fulfill such criteria, while 14.3% were defined as false positives. Among schizophrenics 76.3% achieved this condition while 23.7% were false negatives. The technique's objectivity and ease of application could facilitate the diagnosis of this disease.

Discriminating value of total minor physical anomaly score on the Waldrop scale between patients with bipolar I disorder and normal controls

Minor physical anomalies (MPAs) are slight structural aberrations indicative of abnormal neurodevelopment. Most studies of MPAs in bipolar disorder have yielded limited results. We attempted to assess the potential value of MPAs as a classifying test in the status bipolar I patients vs. normal controls. Sixty one bipolar I patients and 103 controls were evaluated for MPAs using a slightly modified version of the Waldrop scale. The specificity, sensitivity and predictive value of different total MPA (MPA-T) scores were determined. The cut-off MPA-T scores that optimally discriminated patients from controls (exhibiting the most balanced sets of sensitivity, specificity, positive and negative predictive values) were MPA-T ≥ 4 and MPA-T ≥ 5. These values set a "border zone" in which bipolar I patients began to prevail significantly over controls. The latter presented most frequently with MPA-T ≤ 3 and rarely with MPA-T ≥ 6. Bipolar I patients prevailed among outliers (subjects with significantly higher MPA-T scores). Our data establish MPA-T score as a reliable index in distinguishing between bipolar I patients and normal controls and are consistent with the hypothesis of abnormal neurodevelopment in bipolar disorder.