Disentangling negative and positive symptoms in Schizophrenia and Autism Spectrum Disorder

Abstract

Autism Spectrum Disorder (ASD) and Schizophrenia (SZ) share traits, especially in social skills and negative symptoms, and to a lesser degree positive symptoms. Differential diagnosis can be challenging and discerning expressive and experiential negative symptoms may provide knowledge with potential diagnostic and functional relevance that can guide treatment. Two exploratory factor analyses (EFA) were conducted to reveal the underlying dimensions of negative and positive symptoms using the Positive and Negative Syndrome Scale (PANSS), the Scale for the Assessment of Positive Symptoms & Negative Symptoms (SAPS/SANS) and the Autism Diagnostic Observation Schedule-Generic (ADOS-G). Three factors emerged from the negative symptom EFA (70.5% variance): NF1) Expressive Negative; NF2) Experiential Negative; NF3) Preoccupation, Absorption & Expressive Affective Flattening. Three positive factors emerged (68.6% variance): PF1) Hallucinations-Delusions; PF2) Grandiosity; PF3) Thought Disorder-ADOS positive Symptoms. SZ showed higher PF1 scores, and ASD had higher PF3 scores. No differences between groups were observed in the negative factors. Across groups, all negative factors were inversely associated with quality of life. Only NF1 and NF2 and PF1 were detrimentally related to social functioning. A discriminant function analysis using all factors classified correctly 84.4% of participants, with PF1, NF1 followed by NF2 being the best predictors of diagnosis. Expressive negative followed by Experiential negative symptoms are of diagnostic value independent of and beyond SZ-related positive symptoms and are related with detrimental functioning. Findings confirm the need to distinctively target negative symptoms, and specific SZ-related and ASD-related positive symptoms, and especially the use of several assessment tools for diagnostic classification.

1. Introduction

For the majority of patients with a diagnosis of schizophrenia (SZ), course of illness is chronic with persistent positive, negative, and cognitive symptoms that lead to social dysfunction(Couture et al., 2011; McGrath et al., 2008; Saha et al., 2005) . While antipsychotic medication may favorably affect positive symptoms, negative symptoms represent a domain that is difficult to treat effectively (Galderisi et al., 2021; Strauss et al., 2021). Research in schizophrenia has distinguished between experiential (avolition, anhedonia, asociality) and expressive negative symptoms (restricted affect, poverty of speech (alogia) (Blanchard and Cohen, 2006; Reddy et al., 2016; Kring and Barch, 2014) and of their distinct functional correlates, as higher experiential symptoms are associated with worse functioning outcomes (Strauss et al., 2013a).

Current research efforts are underway to understand the overlap in symptom phenotypes between Autism Spectrum Disorder (ASD) and SZ (Jutla et al., 2022; Larson et al., 2017). SZ-related symptoms have been reported in a subset of autistic individuals (Chisholm et al., 2015) at higher rates than nonclinical population (Jutla et al., 2022), and ASD-like traits can be present in SZ (Kincaid et al., 2017; Vita et al., 2020). Traditionally, ASD and SZ have been considered as distinct neuropsychiatric diagnostic entities, showing phenotypical heterogeneity, as SZ characterized by “psychotic” symptoms such as delusions and hallucinations, and ASD by deficits in social-communication and restricted and repetitive behaviors (Jutla et al., 2022), as well as distinct developmental time-course trajectories, as ASD is commonly diagnosed in childhood and SZ in late adolescence or early adulthood (Jutla et al., 2022). However, SZ and ASD share differences in the domains of social communication and interaction (SCI) skills, which are required to meet ASD diagnosis and are frequently present in individuals with SZ with predominant negative symptoms (Jutla et al., 2022; Ribolsi et al., 2022). It is markedly important to examine in depth the differential symptoms, especially in adolescence and youth when a subset of ASD may be susceptible to develop SZ. Typical SZ-related (i.e., hallucinations and delusions) and ASD-related (i.e., in social-communication and restricted and repetitive behaviors) symptoms may be less challenging for diagnoses, but newer studies suggest that a subset of individuals may not be captured by those and hence, a deeper understanding of negative symptoms in these population is needed (Foss-Feig et al., 2019, 2016).

Positive-SZ-related symptoms such as hallucinations and delusions, and positive-ASD-related symptoms such as repetitive stereotypic behaviors have been suggested as the most reliable in differentiating between ASD and SZ (Trevisan et al., 2020). However, these symptoms are not always well-defined (Ahmed et al., 2018; Foss-Feig et al., 2016) and given the heterogeneity within both diagnoses (Abu-Akel et al., 2020; Barneveld et al., 2011; Larson et al., 2020), they may not be sufficient for accurate diagnosis or understanding patients’ functional disability. Thus, it is important to better characterize negative symptoms for accurate differential diagnosis and treatment development.

To our knowledge, no study has examined the specific shared/distinct functional relevance of experiential and expressive negative symptoms in SZ and ASD. Traditionally these symptoms have been assessed with different measures in each disorder (e.g., in ASD: Autism Diagnostic Observation Schedule (ADOS); in SZ: Positive and Negative Syndrome Scale (PANSS)) and the conceptual overlap between these measures may contribute to diagnostic confusion. For example, reduced reciprocal communication and/or shared enjoyment with the examiner in the ADOS may be captured in the PANSS as blunted affect. Thus, it is important to disentangle the distinctive disorder specific contributions of these measures. One recent effort has been in the development of the PANSS Autism Severity Score (PAUSS), which captures ASD-related traits in SZ, and has shown a strong correlation with the ADOS (Kästner et al., 2015), and strong discrimination power between SZ with high and low ASD traits (Deste et al., 2018; Kästner et al., 2015). In another attempt, Trevisan and colleagues (2020) reconceptualized the ADOS items as either positive (presence of atypical behavior) or negative (absence of typical behavior), by following the same conceptual system that is used in the PANSS (Trevisan et al., 2020). They proposed that disorder specific positive symptoms could provide the most differentiation between ASD and SZ. However, their classification of negative symptoms lacked disorder specificity as both groups overlapped in ADOS negative symptoms. They also did not examine their social functioning correlates. Notably, neither the PAUSS nor the work by Trevisan and colleagues (2020) included a separate examination of negative experiential and expressive symptoms, as it is often integrated in SZ research with clinical tools such as the Scale for the Assessment of Negative Symptoms (SANS). Therefore, by conducting a thorough examination of the negative symptoms in these populations using scales developed specifically for either SZ or ASD we will potentially have more in-depth information for differential diagnosis and subsequently for the development and application of distinct vs. transdiagnostic treatment interventions.

The present study aimed to (1) examine the shared symptom phenotype between SZ and ASD in positive symptoms and experiential and expressive negative symptoms by conducting two separate Exploratory Factor Analyses (EFA), with specific items from gold standard measures of ASD and SZ symptom assessments; 2) examine the relation of the positive and negative symptom factors to quality of life (QLS) and social functioning (SFS) in across both groups and separately; and 3) examine the diagnostic power of the factors from the symptom assessments by conducting a discriminatory analysis and assess the specific diagnostic weight of each factor.

2. Materials and Methods

2.1. Participants

Participants were 52 adults with ASD and 48 individuals with Schizophrenia or Schizoaffective Disorder (SZ) or meeting criteria for Psychosis Risk Syndrome. This sample was part of two larger studies conducted at the Olin Neuropsychiatry Research Center (Hartford Hospital, HH) and Yale School of Medicine and approved by their Institutional Review Boards (See Supplementary Materials for details and inclusion criteria of the studies). Participants gave informed written consent following approved procedures before participating in the study. See, Table 1 for the demographic and clinical characteristics of the sample.

Table 1.

Demographic Characteristics and Clinical Variables

	All (N =100)	SZ (n =48)	ASD (n =52)	F/ChiSquare	DoF	p
	n (%)	n ( % )	n ( % )
Gender a
Male	77 (76.2)	34 (69.4)	43(82.7)	2.465	1(1)	.116
Female	24 (23.8)	15 (30.6)	9 (17.3)
Ethnicity a
African American	7 (6.9)	4 (8.2)	3 (5.8)	11.62	4(1)	.02
Asian	2 (2.0)	2 (4.1)	0
Caucasian	68 (67.3)	26 (53.1)	42 (80.8)
Hispanic/Latino	20 (19.8)	13 (26.5)	7 (13.5)
More than 1	4 (4.0)	4 (8.2)	0
	Mean (SD)	Mean (SD)	Mean (SD)
Age (years)	23.91(4.1)	25.81 (3.61)	22.15 (3.75)	24.54	1,98	<0.001
IQ Estimate	105.23 (16.07)	97.96(13.36)	111.94(15.53)	23.10	1,98	<0.001
ADOS Module 4 b *
Communication	3.12 (1.76)	2.75 (2.04)	3.47 (1.37)	2.942	1,95	0.090
Social Interaction	6.09 (3.02)	5.44 (3.60)	6.71 (2.21)	3.624	1,95	0.060
ADOS positive score	1.01 (1.02)	.60 (0.79)	1.39 (1.07)	14.79	1,95	<0.001
ADOS negative score	4.88 (2.12)	4.77 (2.52)	5.00 (1.68)	1.207	1,95	0.275
PANSS c *
Positive	12.80 (4.46)	14.96 (4.66)	10.49 (2.817)	18.591	1,89	0.001
Negative	18.12 (6.58)	19.56 (7.15)	16.58 (5.56)	3.42	1,89	0.067
Cognitive	13.06 (3.79)	14.21 (3.53)	11.84 (3.71)	2.80	1,89	0.097
Hostility	5.08 (1.48)	5.23 (1.65)	4.91 (1.27)	.36	1,89	0.550
Emotion	9.69 (3.55)	10.60 (3.66)	8.71 (3.18)	14.156	1,89	<001
SAPS d,*
Hallucinations	1.15 (1.7)	2.09 (1.87)	0.16 (0.77)	15.5089.	1,87	<0.001
Delusions	1.65 (1.32)	2.17 (1.41)	1.09 (0.96)	164	1,87	0.003
Bizarre Behaviors	0.54 (0.91)	0.64 (0.98)	0.43 (0.81)	0.4530.5	1,87	0.50
Thought Disorder	0.84 (1.14)	0.87 (1.20)	0.80 (1.09)	18	1,87	30.473
SANS e *
Affective Flattening	2.25 (1.54)	2.13 (1.78)	2.37 (1.25)	0.145	1,89	0.704
Alogia	1.54 (1.41)	1.72 (1.48)	1.35 (1.32)	1.04	1,89	0.310
Avolition/Apathy	1.16 (1.33)	1.62 (1.36)	0.70 (1.15)	7.252	1,89	0.008
Anhedonia/Asociality	2.83 (1.48)	3.02 (1.49)	2.63 (1.46)	0.144	1,89	0.705
Social Functioning Scale (SFS) f *	128.03 (23.87)	126.55 (20.84)	129.39 (26.49)	2.682	1,94	0.105
Quality of Life (QLS) g *	76.92 (20.86)	70.44 (19.52)	83.40 (20.32)	3.849	1,92	0.053
Chlorpromazine Equivalents h *	226.08 (396.48)	328.93 (350.57)	131.14 (415.62)	3.2718	1,96	0.057

^*Analysis was conducted with IQ estimate and age as a covariate

^agender and ethnicity included 49 SZ and 52 ASD (N=101)

^bincludes 48 SZ, 51 ASD (N = 99)

^cincludes 48 SZ and 45 ASD (N=93)

^dincludes 47 SZ and 44 ASD

^eincludes 47 SZ and 46 ASD

^fincludes 48 SZ and 48 ASD

^gincludes 47 SZ and 51 ASD

^hincludes 48 SZ and 52 ASD

2.2. Measures

Refer to Supplementary Materials 2A for full description of Measures. Participants received an interview inquiring about demographics, psychiatric history, and medication, and the following measures were used.

Intellectual Quotient (IQ) and Symptom Measures

Wechsler Scale of Adult Intelligence-III (WAIS-III) (Sattler and Ryan, 1999; Wechsler, 1997): Full Scale IQ was estimated using the Vocabulary and Block Design Subtest subtests of the WAIS-III.

Structured Clinical Interview for DSM-IV (SCID) (First, 1997) was used operationalize criteria for diagnostic classification based on the DSM-IV (APA, 1994).

Positive and Negative Syndrome Scale (PANSS) (Kay et al., 1987): A semi-structured questionnaire that assesses psychosis-related symptoms. We calculated 5 component scores: Positive, Negative, Cognitive, Hostility, and Emotional Discomfort components (Bell et al., 1994).

Autism Diagnostic Observation Schedule-Generic -Module 4 (ADOS-G) (Lord et al., 2000): This semi-structured observational tool assesses ASD-related symptoms. Besides scores with standard procedures, additional scores were calculated following the procedures described by Trevisan et al (2020), which categorized ADOS items as present symptoms of atypical behavior, or “positive”, and the absence of typical symptoms as “negative”.

The Scale for the Assessment of Positive Symptoms & Negative Symptoms (SAPS & SANS)(Andreasen, 1989; 1984): SAPS addresses four types of positive symptoms: a) hallucinations; b) delusions; c) bizarre behavior and d) positive formal thought disorder. SANS examines 5 types of negative symptoms, (a) affective flattening or blunting; (b) alogia; (c) avolition-apathy; (d) anhedonia-asociality; (e) problems in attention.

Social functioning measures

Heinrichs Quality of Life Scale (QLS) (Heinrichs et al., 1984): Semi-structured interview assessing functioning into four domains of function: interpersonal relations, intrapsychic foundations, instrumental role function, and common objects and activities.

Social Functioning Scale (SFS) (Birchwood et al., 1990): It measures social functioning in seven functioning areas: social engagement/withdrawal, interpersonal behavior, performance skills for independent living, recreation, prosocial activities, ability to perform skills for independent living, employment/occupation.

2.3. Data Analysis

Analyses were conducted in three phases. Initially, groups were compared on demographic, chlorpromazine equivalents of anti-psychotic medications, and IQ estimate variables using Chi-square tests for categorial variables and analysis of variance (ANOVA) for continuous variables. Due to between group differences in age and IQ estimate, they were subsequently included as covariates in all models. Next, groups were characterized and compared on clinical symptoms, social functioning and quality of life measures using analysis of covariance (ANCOVA). Subsequently, two exploratory factor analyses (EFAs) were conducted to capture dimensional constructs within positive and negative symptoms using the following measures and items:

1^st EFA: Negative Symptoms (13 items)

• PANSS (Bell et al., 1994; Kay et al., 1987) - 8 items comprising the negative symptom subscale of the 5-factor structure (n1+n2+n3+n4+n6+g7+g13+g15)

• SANS (Andreasen, 1989) - 4 items: 2 Expressive negative subscales: (a) affective flattening or blunting and (b) alogia; and 2 Experiential negative subscales: (c) avolition-apathy (d) anhedonia-asociality.

• ADOS-G (Lord et al., 2000) - Negative Symptoms Score following the scale developed by Trevisan et al.(2020).

2^nd EFA: Positive Symptoms EFA (11 items):

• PANSS (Bell et al., 1994; Kay et al., 1987). 6 items comprising the positive symptom component from the 5-factor structure (p1+p3+p5+p6+g1+g9)

• SAPS (Andreasen, 1989, 1984) −4 items- from the positive subscales: (a) hallucinations; (b) delusions; (c) bizarre behaviors; (d) positive formal thought disorder

• ADOS-G (Lord et al., 2000) Positive Symptoms Score following the scale designed by Trevisan et al (2020).

Before performing each EFA, items that correlated ≥0.7 were removed to avoid initial multicollinearity. The EFAs were performed using an oblique rotation (direct oblimin) and Kaiser normalization (Field, 2018). For each EFA, eigenvalues >1 served as the selection criteria and scree plots were examined to confirm factor selection (Cattell, 1966). Items with factor loadings of ≥ 0.3 were considered representative of a given factor. Items with communalities less than 0.2 were dropped (Costello and Osborne, 2005). Final factors for each EFA were compared between groups using ANCOVA controlling for age and IQ. Correlation analyses were used to examine associations between each factor and measures of social functioning (SFS) and quality of life (QLS) within groups, and partialing for age and IQ in across groups. Fisher r-to-z tests were conducted to test for differences in correlations between groups. Lastly, the diagnostic predictive value of the factors was assessed using a discriminant analysis with diagnosis as the DV and the final factors as predictors.

3. Results

3.1. Sample Characterization

Demographic variables, estimates of IQ, and clinical and social functioning scores for each diagnostic group are shown in Table 1. SZ subjects were significantly older compared to ASD subjects while the gender distributions were similar between the groups. The SZ group was more ethnically diverse than the ASD group, although both groups were primarily comprised of Caucasian males. Additionally, the ASD group had significantly higher IQ scores compared to SZ. After controlling for IQ scores and age groups did not differ in chlorpromazine equivalent doses. As expected, higher ADOS positive scores were observed among ASD subjects. The SZ group showed significantly higher PANSS 5-factor Positive and Emotion scores, greater SAPS hallucinations and delusions subscale scores, and greater SANS avolition/apathy subscale. No group differences were observed in quality of life or social functioning.

3.2. Exploratory Factor Analyses

3.2.1. 1^st EFA: Negative Symptoms

The EFA with the 13 negative symptoms items revealed 3 independent factors accounting for 62.2% of total variance: NF1) Expressive Negative Symptoms (41.6%); NF2) Experiential Negative Symptoms (11.2%); NF3) Preoccupation, Absorption & Expressive Affective Flattening (9.47%). Two items (PANSS G13: Disturbance of volition, SANSS: Global rating of avolition/apathy) were dropped due to low communalities (<0.2) and factor loadings <0.3. As shown in Table 2, the final EFA with 11 items produced 3 independent factors, explaining 70.5% of total variance: NF1) Expressive Negative Symptoms (46.7%); NF2) Experiential Negative Symptoms (12.7%); NF3) Preoccupation, Absorption & Expressive Affective Flattening (11.1%).

Table 2.

Factor Loadings of the 1^st EFA Negative Symptom Measures in a 3 Factor Solution after Varimax Rotation (Total variance explained of 70.48%)

	Components
	1 Expressive Negative Symptoms (46.65% variance)	2 Experiential Negative Symptoms (12.73% variance)	3 Preoccupation and Absorption & Expressive Affective Flattening (11.10% variance)
ADOS- Negative Symptom Score	.404	.280	.121
PANSS -N1 Blunted Affect	.705	−.021	.299
PANSS-N2 Emotional Withdrawal	.267	.588	−.045
PANSS-N3 Poor Rapport	.732	.103	.033
PANSS-N4 Passive -Apathetic Social Withdrawal	.021	.930	−.170
PANSS-N6 Lack of Spontaneity and Flow of Conversation	.995	−.074	−.123
PANSS-G7 Motor Retardation	.725	−.009	−.150
PANSS-G15 Preoccupation	−.081	−.027	.478
SANS- Affective Flattening or Blunting	.352	.059	.572
SANS- Alogia	.591	.241	.128
SANS- Anhedonia/Asociality	−.109	.844	.106

Note: These are final 11 item loadings of the Negative EFA after removing two items that originally had low communalities and low factor loadings, the item “G13 Disturbance of volition: and the item “SANSS: Global rating of avolition/apathy”.

3.2.2. 2^nd EFA: Positive Symptoms

The items SAPS H7 (Global Rating of Hallucinations) and SAPS D1 (Global Rating of Delusions) were excluded from the analysis due to their high correlation with their PANSS counterparts (r=0.9 and r=0.7, respectively). After oblique rotation the remaining 9 items produced a solution which included two items with low communalities and factor loadings (PANSS G1: Somatic Concern, SAPS: Bizarre behaviors) and thus dropped. As shown in table 3, the final EFA among the remaining 7 items a threefactor solution that accounted for 68.6% of total variance: PF1) Hallucinations and Delusions Symptoms (36.2%); PF2) Grandiosity (18.1%); and PF3) Thought Disorder and ADOS Positive Symptoms (14.3%). Although PF2 and PF3 each had only one item with a high loading the factors were retained given their differential explanatory power (PANSS P5: Grandiosity, 0.657 loading on PF2, and P9: Positive Thought Disorder 0.496 loading on PF3). ADOS positive symptoms loaded negatively on PF1 (−0.431) but positively on PF3 (0.422).

Table 3.

Factor Loadings of the 2^nd EFA Positive Symptom Measures in a 2 Factor Solution after Varimax Rotation (Total variance explained of 55.34%)

	Components
	1 Hallucinations- Delusions Symptoms (36.184% variance)	2 Grandiosity (18.088% variance)	3 Thought Disorder and ADOS Positive Symptoms (14.331% variance)
ADOS- Positive Symptom Score	−.431	.085	.422
PANSS -P1 Delusions.	.742	−.050	−.028
PANSS- P3 Hallucinatory Behavior	.834	−.047	−.035
PANSS-P5 Grandiosity	.026	.657	−.086
PANSS- P6-Suspiciousness Persecution	.508	.013	.060
PANSS G9-Unusual Thought Content	.491	.384	.125
SAPS- Positive Formal Thought Disorder	.237	-.205	.496

Note: These are final 7 item loadings of the Positive EFA after removing two items with high correlation (SAPS. Global Ratings of Hallucinations and SAPS. Global Rating of Delusions) and one item that showed a communalities <0.2 and did not load in any factor, the PANSS item “G1: Somatic Concern” and the SAPS item- Bizarre Behavior

3.3. Group Differences between the Outcome factors of the EFAs

No group differences were observed for any of the negative factors: NF1: Expressive Negative Symptom, F(_1,87)= 2.70; p=0.104); NF2: Experiential Negative Symptoms factors (F(_1,87)= 0.25; p=0.62); NF3:Preoccupation and Absorption & Expressive Affective Flattening (F(_1,87)=0.01; p=0.99). However, significantly higher PF1: Hallucinations-Delusions scores were observed in SZ compared to ASD (F(_1,86) = 23.5, p < 0.001) whereas ASD had higher scores for PF3:Thought Disorder and ADOS Positive Symptoms (F(_1,86)=5.9; p=0.017). No differences between groups were observed for PF2: Grandiosity (F(1,86)= 0.25; p=0.88).

3.4. Bivariate Correlations between Outcome factors of the EFAs and Social functioning

While controlling for age and IQ negative associations were observed between NF1) Expressive Negative Symptoms factor and both QOL (SZ+ASD: r = −0.49; p=0.001; SZ: r= −0.50; p<0.001; ASD: r= −0.56; p<0.001) and SFS (SZ+ASD: r = −0.22; p=0.04; SZ: r= −0.16; p=0.29; ASD: r= −0.32; p<0.037). Similarly, strong negative association were observed between NF2) Experiential Negative Symptoms and both QLS (SZ+ASD: r = −0.65, p<0.001; SZ: r= −0.67, p<0.001; ASD: r= −0.70; p<0.001) and SFS (SZ+ASD: r = −0.48; p<0.001; SZ: r= −0.39; p=0.008; ASD: r= −0.55; p<0.001. Similarly, NF3) Preoccupation and Absorption & Expressive Affective Flattening correlated negatively with QLS (SZ+ASD: r = −0.28, p=0.01; SZ: r= −0.24, p<0.11; ASD: r= −0.31; p=0.048, but not with SFS. However, PF1) Hallucinations-Delusions factor was negatively correlated with SFS (SZ+ASD: r = −0.23; p=0.03; SZ: r= −0.17; p=0.27; ASD: r= −0.24; p=0.13. None of the above correlations differed between groups.

3.5. Discriminant Analysis

The value of the estimated single discriminant function was significantly different between SZ and ASD (chi-square = 64.9, df = 6, p<0.001). Correlations between the predictor variables and the standardized canonical discriminative function revealed that the best predictors of diagnosis were PF1 (Hallucinations-Delusions, coefficient = 1.195) and NF1 (Expressive Negative symptoms, coefficient = .836), followed by NF2 (Experiential Negative Symptoms; coefficient = −0.520) suggesting that patients with higher number of positive SZ symptoms and expressive negative symptoms, but lower experiential negative symptoms were more likely to be diagnosed with SZ. The other three factors showed weaker negative correlations (< −0.3). The discriminant function successfully predicted diagnosis for 84.4% of cases, with accurate predictions of 83.0% of SZ patients and 86.0% of ASD patients (Table 4).

Table 4.

Discriminant Analysis with diagnosis as the DV and 5 outcome factors of the EFAs as predictors

Classification Resultsa
			Predicted Group Membership		Total
			Schizophrenia	Autism
Original	Count	Schizophrenia	39	8	47
		Autism	6	37	43
	%	Schizophrenia	83	17	100
		Autism	14	86	100

^a.84.4% of original grouped cases correctly classified.

4. Discussion

The present study revealed: 1) Three negative symptom and three positive symptom factors across a sample of adults diagnosed with either SZ or ASD: NF1: Expressive Negative Symptoms, NF2: Experiential Negative Symptoms and NF3: Preoccupation, Absorption & Expressive Affective Flattening factors and PF1: Hallucinations-Delusions, PF2: Grandiosity and PF3: Thought Disorder and ADOS Positive Symptoms; 2) SZ showed higher significant scores in PF1 and ASD in PF3; 3) QLS was significantly inversely related with the NF1,NF2 and NF3factors, and social functioning was only negatively related with NF1, NF2 and PF1 across both groups; 4) The six factors successfully predicted diagnosis for 84.4% of the cases, revealing a higher prediction accuracy for ASD than for SZ, with Hallucinations-Delusions (PF1), Expressive negative symptoms (NF1), followed by Experiential negative symptoms (NF2)as the best predictors of diagnosis in which the higher the PF1 and NF1 symptoms the more likely a patient would be diagnosed with SZ, and the higher the NF2 with ASD

The Negative Symptom EFA with the originally selected 13 items from the PANSS, SANS and ADOS, initially revealed three factors, however, two of those items, G13: Disturbance of volition and SANS-item: Global rating of avolition/apathy, showed very low factor loadings (all<0.3), and did not produce an additional factor. Interestingly, the SANS item: Global rating of avolition/apathy was the only SANS item that differed between groups, in which the SZ group showed significantly higher scores. Current literature proposes “amotivation/avolition” as the core central primary construct of negative symptoms in SZ, from which the other ones would derive (Foussias and Remington, 2010; Marder and Galderisi, 2017; Strauss et al., 2021, 2020), however, in ASD, social motivation deficits may be attributed to comorbid anxiety (Bagg et al., 2023; Swain et al., 2015), and therefore could be considered “secondary” symptoms. This possible etiological distinction on “amotivation/avolition” in SZ and ASD could be the source of future studies examining new differential diagnosis methods. In the subsequent EFA without them, it produced three robust factors that explained 70.48% of the variance. From them, NF1) Expressive Negative Symptoms, explained most of the variance, comprising items capturing lack of expressive behaviors, including facial and verbal expression, social initiative, and slowness in movements. It was not surprising to find that the ADOS negative symptom item loaded more heavily in NF1, as it captures the decrease in SCI-related behaviors, such as lack of facial expression directed to examiner. These behaviors (or lack thereof) are naturally overtly expressed which are comparable to the “expressive negative symptoms” items from the PANSS and the SANS item found in this EFA. As to Trevisan et al.,(2020)both groups shared deficits in NF1 and may explain the large percentage of SZ (43.59%) with “false positives” (i.e. misdiagnosed as SD) found in their study by using the ADOS, thus, confirming that the ADOS alone may not be sufficient for differential diagnosis. Similarly, NF2: Experiential Negative Symptom did not differ between groups. The fact that NF2 did not differ between groups illustrated that ASD also showed deficits in experiential symptoms measured with scales developed for psychosis disorders. Research has shown that individuals with ASD have extensive deficits in the processing of social and non-social rewards (Aldridge-Waddon et al., 2020; Clements et al., 2018; Keifer et al., 2021) which may be leading the general motivational abnormalities observed in ASD and impacting social development (Clements et al., 2018; Keifer et al., 2021). This reduced ability to anticipate rewards (social and non-social) has been proposed as a potential transdiagnostic phenotype as it is shown in other psychiatric disorders besides ASD and SZ, such as depression (Aldridge-Waddon et al., 2020; Keifer et al., 2021). Deficits in social and emotional withdrawal, avolition and anticipatory anhedonia are hallmark symptoms of SZ (Edwards et al., 2015; Reddy et al., 2016; Strauss et al., 2021) and largely related to greater deficits in social functioning compared to deficits in expressive symptoms (Reddy et al., 2016; Gard et al., 2007; Strauss et al., 2021, 2013a). In our study, both the NF1 (Expressive Negative Symptoms) and NF2 (Experiential Negative Symptoms) factors showed negative correlations with QLS and SFS, however, in support of supports the current literature, the NF2 showed higher effect sizes. The NF3 (Preoccupation and Absorption & Expressive Affective Flattening) was also negatively correlated with QLS but showed the weakest effect size and had no association with SFS.

While NF1 and NF2 reflected the two main negative symptom domains reported previously in SZ: the expressive and the experiential domain without avolition (Blanchard and Cohen, 2006; Kirkpatrick et al., 2006; Kirkpatrick and Fischer, 2006; Kring and Barch, 2014; Messinger et al., 2011), the current EFA, conducted with SZ and ASD samples, revealed a third negative factor, NF3: Preoccupation, Absorption & Expressive Affective Flattening. This factor was comprised of the PANSS item G15: Preoccupation that measures observable behaviors of self-absorption, along with the SANS expressive affective flattening item and was the only negative factor in which ASD showed higher factor scores than the SZ, although not significant. It was surprising that the items of this factor did not load in NF1 as they also captured expressive negative deficits, however, NF3 distinctly reflected “self-centered/ egotistical” responses, combining symptoms of internal preoccupation and self-absorption with disregard to reality with a generalized affective blunting and emotional inappropriateness. Similarly, the PAUSS, which aimed to capture ASD symptoms in SZ, also included preoccupation as one of the symptoms (Kästner et al., 2015) thus, supporting this item as being more pronounced in ASD. Although NF3 did not correlate with social functioning and was not the strongest predictor of diagnosis, it provided insightful information on the negative symptoms more associated with ASD compared to SZ.

The Positive Symptom EFA revealed three factors with two distinct SZ-related and ASD-related positive symptom factors. The PF1: Hallucinations-Delusions reflected psychotic symptoms central to the SZ diagnosis, such as hallucinations, delusions, paranoid ideation, and unusual thought content, additionally the ADOS Positive Symptom score, which reflects ASD-related symptoms, was heavily inversely loaded with this factor. On the other hand, the PF3: Thought Disorder and ADOS Positive Symptoms revealed key symptoms of ASD. It was supported by the thought disorder item of the SAPS which measures behaviors that prevent effective communication, such as tangentiality and the ADOS- positive item loaded also heavily on this factor, but positively, which captures observable ASD symptoms, such as repetitive, inappropriate behaviors and unusual sensory interests, De Crescendo et al. (2019) revealed that tangential thought and focus on favorite subjects which are commonly seen in ASD, can be also conceptualized as thought disorder in SZ. Additionally, Deste et al.(2021) described that only a minority of individuals with SZ in their study did not show any ASD symptoms, and the ones with higher ASD symptoms had worse real-world outcomes.

Both the PF1 and PF3 are consistent with the core symptoms of SZ-related and ASD-related, respectively demonstrated by Konstantareas & Hewitt (2001), Trevisan et al. (2020), and Han et al (2022). Our results support the conceptual distinction of the SZ and ASD-related positive symptoms shown by these two positive factors (PF1 and PF3) following DSM-5 (American Psychiatric Association, 2013) conventions, and both differed between groups, but only PF1 showed the highest predictive power of diagnosis. (De Crescenzo et al., 2019; Deste et al., 2021, 2020b, 2020a)

The PF2 (Grandiosity) did not differ between groups, as several studies have described the overlap in symptoms between ASD and personality disorders (PD), including narcissistic PD (Broglia et al., 2024; Vannucchi et al., 2014). Broglia et al., 2024 described the high incidence of vulnerable narcissism in ASD compared to grandiose narcissism, however, given that the present study used the PANSS G5 item to capture “Grandiosity” as a subtype of delusion in SZ, it may not have captured the subtlety between the types of narcissism, but notwithstanding, was able to capture symptoms of boastfulness, exaggerated selfopinion, unrealistic feelings of superiority to others, increase of self-importance and delusions if great abilities, wealth, knowledge, fame, power or moral righteousness, which may have been equally visible in ASD as in SZ.

Surprisingly, in our study, none of these positive factors correlate with QLS or SFS. Importantly, the symptoms reflected by PF1, PF2, and PF3 may be more responsive to pharmacological interventions and thus may not interfere with functioning as much as negative symptoms as they are treated (Fusar-Poli et al., 2015; Strauss et al., 2021).

Finally, the discriminant analysis revealed that these five factors can successfully predict diagnosis in 84.4% of the cases, revealing that Hallucinations-Delusions (PF1), and Expressive Negative Symptoms (NF1) followed by Experiential Negative Symptoms (NF2) are the best predictors of diagnosis. Given the strong predictive power of these factors with symptoms only, they provide potential framework to further research underlying mechanism in ASD and SZ by examining their relationship with biological markers, such as neural network architecture. These findings confirm that SZ-related positive symptoms and Experiential Negative symptoms measured by traditional SZ measures such as the PANSS/SAPS may provide the highest predictive power to distinguish SZ from ASD, and Expressive Negative Symptoms assessed with PANSS/SAPS and ASD-related positive symptoms as measured by the ADOS may be better at distinguishing ASD from SZ.

Certain limitations of this study could be addressed in future research. We did not account for the role of social cognition in distinguishing these groups. SZ and ASD share deficits in social cognition (Corbera et al., 2021; Oliver et al., 2021; Sasson et al., 2011) and their performance in this domain may also assist in diagnostic classification (Jutla et al., 2022; Sasson et al., 2011)Additionally, other studies should also explore functioning using more proxi-measures of functional capacity such as performancebased measures (Deste et al., 2020b). Future research could also extend the findings by using newer measures of negative symptoms such as the Clinical Assessment Interview for Negative Symptoms (CAINS) (Kring et al., 2013)and the Brief Negative Symptoms Scale (BNSS) (Kirkpatrick et al., 2011). In addition, the group sizes were not equal, and the ASD had higher IQ estimate scores and was younger than the SZ and we subsequently adjusted for both . The samples also were not uniformly racial and ethnically diverse, as they were predominantly Caucasian, especially in the ASD group, and it may have impacted the results and their generalizability. Finally, both groups were taking medication, and we cannot rule out their influence on symptoms.

5. Conclusions

Despite these limitations, the present study enhanced our understanding of the shared symptom phenotype between SZ and ASD, especially in the negative symptom dimension. We found that not only SZ-related positive symptoms can aid in diagnostic classification but also expressive and experiential negative symptoms. Negative symptoms (expressive and experiential) were strong predictors of functioning with the experiential ones being the strongest in both groups, surpassing the functional relevance of positive SZ and ASD-related symptoms. Overall, these results have important implications confirming the need to target negative symptoms, especially the experiential ones, with cross-diagnostic interventions for the improvement of functioning. Furthermore, they encourage the use of more than one assessment tool for diagnostic classification, as one symptom assessment tool developed for a specific diagnostic group may not capture the symptom nuances required for differentiation.