Aberrant Hippocampal Connectivity in Unmedicated Patients With Schizophrenia and Effects of Antipsychotic Medication: A Longitudinal Resting State Functional MRI Study

Abstract

To better characterize hippocampal pathophysiology in schizophrenia, we conducted a longitudinal study evaluating hippocampal functional connectivity during resting state, using seeds prescribed in its anterior and posterior regions. We enrolled 34 unmedicated patients with schizophrenia or schizoaffective disorder (SZ) and 34 matched healthy controls. SZ were scanned while off medication, then were treated with risperidone for 6 weeks and re-scanned (n = 22). Group differences in connectivity, as well as changes in connectivity over time, were assessed on the group’s participant level functional connectivity maps. We found significant dysconnectivity with anterior and posterior hippocampal seeds in unmedicated SZ. Baseline connectivity between the hippocampus and anterior cingulate cortex, caudate nucleus, auditory cortex and calcarine sulcus in SZ predicted subsequent response to antipsychotic medications. These same regions demonstrated changes over the 6-week treatment trial that were correlated with symptomatic improvement. Our findings implicate several neural networks relevant to clinical improvement with antipsychotic medications.

Introduction

Post-mortem and in vivo neuroimaging studies suggest extensive hippocampal pathology in schizophrenia. Findings include volume loss,¹ decreased neuronal integrity,² and aberrant glutamatergic and ɣ-Aminobutyric-acid (GABA)-ergic signaling.^3,4 On a functional level, excess hippocampal drive, as evidenced by elevated regional cerebral blood flow (rCBF),⁵ blood volume,⁶ and intrinsic activity⁷ has been reported.

Imaging studies have shown abnormal hippocampal modulation during episodic memory encoding^8–10 and retrieval,^8–13 novel picture encoding,¹⁴ and relational learning in schizophrenia.¹⁵ Measuring synchronicity in low frequency fluctuations of the BOLD (Blood Oxygen Level Dependent) signal between brain voxels¹⁶ during resting state, we recently described decreased connectivity between the left hippocampus and bilateral precuneus in unmedicated patients compared to controls.¹⁷ In medicated patients, aberrant hippocampal resting state connectivity with the parahippocampal gyrus, medial prefrontal cortex (MPFC), anterior cingulate cortex (ACC), posterior cingulate cortex (PCC), striatum, and cerebellum have been reported,^18–20 albeit directionality of abnormalities (increase/ decrease/ mixed increase and decrease) appears inconsistent. Rodent and primate studies suggest differential projections to cortical and subcortical areas along the longitudinal axis of the hippocampal formation,²¹ which is further supported by human neuroimaging studies that have identified different patterns of functional connectivity along the anterior–posterior axis.^22,23 Additionally, studies have also suggested hemispheric lateralization abnormalities in schizophrenia,²⁴ including in the hippocampus.^25–27 It is plausible that discrepancies in the literature may be partially attributable to differences in the definition of the region of interest, underscoring the importance of considering anatomical and functional heterogeneity when examining hippocampal pathology.

Because aberrant mesolimbic dopamine function contributing to positive symptoms may be secondary to excessive hippocampal drive,⁴ and positive symptom severity has been associated with hippocampal and ACC rCBF in unmedicated patients,²⁸ investigation of hippocampal circuitry could advance our understanding of antipsychotic drug action. We previously reported elevated hippocampal rCBF and glutamate in unmedicated patients that appear to be attenuated with antipsychotic treatment.^29,30 Importantly, a reduction in hippocampal rCBF as well as change in functional connectivity to the MCFC with medication were associated with good clinical outcomes.^29,31 We also demonstrated that treatment response was predicted by resting state connectivity between the ventral tegmental area/midbrain and the dorsal ACC as well as the default mode network, including the hippocampus, in unmedicated patients.³²

In the present study, we evaluated hippocampal functional connectivity during resting state, using seeds prescribed in its anterior and posterior regions. We obtained scans when patients were off medication and after 6 weeks of treatment with risperidone, a second-generation antipsychotic. We hypothesized to find a mixed pattern of increased and decreased connectivity from anterior and posterior seeds to the parahippocampus, MPFC, ACC, PCC, precuneus, striatum, and cerebellum that would be modulated by antipsychotic treatment. We also hypothesized that baseline connectivity patterns within cortico-limbic-striatal circuits would relate to clinical response after 6 weeks of treatment,^32,33 and that connectivity with the MPFC, ACC, and striatum would change as a function of treatment response.^29,32,34

Methods

Unmedicated patients with schizophrenia or schizoaffective disorder (SZ) were recruited from the inpatient unit, outpatient clinics, and emergency room at the University of Alabama at Birmingham (UAB). Healthy controls (HC) matched on age, sex, parental occupation, and smoking were recruited by advertisements. Approval was obtained from the UAB Institutional Review Board. Written informed consent was obtained prior to enrolment after subjects were deemed competent to provide consent.³⁵

Diagnoses were established by review of medical records, the Diagnostic Interview for Genetic Studies (DIGS), and consensus of 2 board certified psychiatrists (A.C.L. and N.V.K.).³⁶ The Brief Psychiatric Rating Scale (BPRS) was used to assess symptom severity.³⁷

Subjects were excluded if they had major neurological or medical conditions, history of head trauma with loss of consciousness, substance use disorders (excluding nicotine) within 6 months of imaging, were prescribed medications known to affect brain function, were pregnant or breastfeeding, or had magnetic resonance imaging (MRI) contraindications. Controls with a personal or family history in a first-degree relative of an Axis I disorder were excluded.

Subjects who were medication naïve or had been off antipsychotic medications for at least 2 weeks (determined by self report) were enrolled in a 6-week trial of risperidone using a flexible dosing regimen. Resting state scans were obtained prior to treatment (off medication), and after 6 weeks of treatment. Nineteen HC were scanned twice, 6 weeks apart, to examine variability of imaging data. Medication was managed by ACL and NVK. Risperidone was started at 1–3 milligrams and titrated in 1–2 milligram increments, dosing was based on therapeutic and side effects. Use of concomitant medications was permitted as clinically indicated (benztropine for 12 subjects, trazodone for 2, mirtazapine, amitriptyline, and valproic acid for 1 subject each). Compliance was monitored with pill counts at each weekly visit. Patients reported adherence to antipsychotic medications, with only few having missed an occasional dose.

Of the 68 subjects enrolled, 1 HC was excluded from baseline analysis because of poor scan quality. Six SZ dropped out of the study prior to the second scan. Additionally, no resting state data was available for 6 SZ subjects at week 6 (resting state scans not obtained for 5 subjects, poor scan quality in 1 subject), leaving data for 22 SZ at week 6. To maximize signal to noise ratio, we also excluded individual connectivity maps of seeds with EPI dropout. All subjects contributed maps from at least 1 seed (left anterior: SZ—22 baseline, 18 week 6, HC—21 baseline, 12 week 6; right anterior: SZ—30 baseline, 17 week 6, HC—24 baseline, 12 week 6; left posterior: SZ—34 baseline, 21 week 6, HC—33 baseline, 14 week 6; right posterior: SZ—34 baseline, 22 week 6, HC—30 baseline, 14 week 6). Resting state data from 22 SZ and 22 HC were included in earlier reports.^17,32

MRI Acquisition

Imaging was performed on a 3T scanner (Magnetom Allegra, Siemens). Resting state scans were acquired during a 5-minute gradient recalled echo-planar imaging sequence (repetition time/echo time [TR/TE]= 2000/30ms, flip angle = 70°, field of view = 192×192mm², 64×64 matrix, 6-mm slice thickness, 1mm gap, 30 axial slices, 150 acquisitions). During the scan, subjects were instructed to keep their eyes open and stare passively ahead. High-resolution structural scans were acquired using the T1-weighted magnetization prepared rapid acquisition gradient-echo sequence (TR/TE/inversion time [TI] = 2300/3.93/1100ms, flip angle = 12°, 256×256 matrix, 1-mm isotropic voxels). All scans were reviewed by a neuroradiologist.

Functional Network Analysis

Using the SPM8, resting state data were slice timing corrected, realigned using rigid-body motion transforms, co-registered to the high-resolution structural scan, normalized to Montreal Neurologic Institute (MNI) space and spatially smoothed employing the diffeomorphic anatomical registration using exponentiated lie algebra algorithm (DARTEL)³⁸ with a 6-mm at full-width half-maximum 3-dimensional Gaussian kernel. To assess motion effects on connectivity data, we calculated the mean absolute displacement of the brain from one timeframe to the next for each subject. There were no significant differences in mean absolute displacement between HC and unmedicated SZ, or between time points in SZ (all P > .05).

A nuisance regression was conducted using the 6 motion parameters identified during the realignment step and their first derivatives as regressors to reduce the effect of movement. A step-wise data scrubbing procedure followed to reduce the effects of time points severely contaminated by motion.³⁹ Frame wise displacements from each time point, i, to the next (FD_i) were computed from the 6 realignment parameters. Time points with FD_i > 0.5mm were considered to be severely contaminated. These contaminated time points were first interpolated prior to bandpass filtering of the data (0.009 < f< 0.08 Hz),⁴⁰ and then excluded from subsequent analyses.⁴¹ Following data “scrubbing,” a principle component analysis was used to extract the components of white matter and cerebral spinal fluid necessary to explain 90% of the signal variance from those regions. Extracted components were used as regressors in a second nuisance regression.⁴²

Spherical seeds with 6mm diameter were placed at the following MNI coordinates: ±30/−10/−20 for the anterior hippocampus and ±32/−34/−6 for the posterior hippocampus. Coordinates were chosen based on Kahn et al²² and in agreement with Poppenk’s recommendations for coordinate-based functional hippocampus segmentation.⁴³ Visual inspection confirmed seed regions were placed centrally in gray matter.

The first eigenvariate of the BOLD time series from each region was extracted and correlated to the time series of all other voxels in the brain to produce functional connectivity maps (units of Pearson’s r correlation). Group differences were assessed using 2-sample t tests and paired-sample t tests for repeated measures within groups. Analyses were limited (masked) to include only regions that were recruited by at least one of the groups. These masks were computed by thresholding (P _FDR < .05), binarizing and then combining each within-group analysis (both times points in SZ) producing masks of the combined connected areas. These masks were entered as inclusive explicit masks in between-group and paired-sample analyses. All analyses were corrected for multiple comparisons using the false discovery rate (FDR) and are reported at P _FDR < .01 for main effects within-group and at P _FDR < .05 for between-group and paired-sample analyses, unless specified otherwise.

Regression analyses were performed to identify, in unmedicated SZ, regions of hippocampal connectivity correlated with eventual treatment response ([BPRS positive score at baseline − BPRS positive score at 6 weeks/ BPRS positive score at baseline] × −100). For each seed, we also created connectivity change maps between baseline and week 6. Change maps were entered into an additional regression analysis using treatment response as the independent variable; these analyses were limited to areas correlated with the response at baseline (small volume correction at P < .05). As a post hoc analysis, we replaced treatment response with percent reduction in hallucinations as independent variable to examine to which extent reductions of hallucinations explained findings.

In an exploratory fashion, we examined the relationship between functional connectivity and symptom burden in SZ at baseline using a general linear model with BPRS positive and negative symptom scores as regressors to examine if any associations unique to symptom dimension exist.⁴⁴

Results

Demographics and Clinical Variables

There were no significant differences between HC and SZ in age, sex, parental socioeconomic status, or smoking (table 1). In SZ, BPRS scores significantly decreased from baseline to week 6 of treatment. The average dose of risperidone at that time was 4.36±1.45mg.

Table 1.

Demographics and Clinical Measures^a

	SZ (n = 34)	HC (n = 34)	t/X2	P value
Gender (% male)	67.6	67.6	0.000	1.0
Age	32.38 (10.43)	32.00 (9.02)	−0.382	.87
Parental occupationb	7.26 (6.39)	6.21 (4.35)	−0.80	.43
Smoking status (% smokers)	76.5	64.7	1.133	.42
Smoking (packs per day)	0.59 (0.53)	0.65 (0.60)	0.447	.66
Diagnosis
Schizophrenia	31
Schizoaffective disorder	3
Illness characteristics
Illness duration (y)	9.59 (9.94)
First episode	12
Prior antipsychotic treatment
Antipsychotic naive	17
Antipsychotic free interval (mo)	23.08 (44.41)
BPRSc
Total
Baseline	48.29 (9.38)
Week 6d	30.57 (8.47)
Positive
Baseline	9.53 (3.04)
Week 6d	4.86 (2.38)
Negative
Baseline	6.79 (2.51)
Week 6d	5.39 (2.42)

Note: SZ, schizophrenia; HC, healthy controls.

^aMean (SD) unless indicated otherwise.

^bRanks determined from Diagnostic Interview for Genetic Studies (1–18 scale); higher rank (lower numerical value) corresponds to higher socioeconomic status.

^cBrief Psychiatric Rating Scale (BPRS) (1–7 scale); positive (conceptual disorganization, hallucinatory behavior, and unusual thought content); negative (emotional withdrawal, motor retardation, and blunted affect).

^d n = 28.

Imaging Results

Functional Connectivity in HC.

In HC, all 4 seed regions showed unique connectivity patterns (supplementary table 1, supplementary figure 1). In HC scanned twice, connectivity was unchanged in 3 seeds, but left posterior hippocampus connectivity increased in 2 clusters (t = 5.82, P _FDR < .05, x = −10, y = −55, z = 23; t = 5.17, P _FDR < .05, x = 34, y = 30, z = −4).

Functional Connectivity in Unmedicated Patients.

In unmedicated SZ compared to HC we found widespread aberrant connectivity with all 4 seeds. Increased connectivity in unmedicated SZ was found between the left anterior hippocampus and MPFC, ACC, and caudate as well as the left posterior hippocampus and right parahippocampus, middle temporal gyrus, inferior parietal cortex, precuneus, cuneus, lingual gyrus, angular gyrus, and cerebellum (figure 1, supplementary table 2). Connectivity was decreased between the right anterior hippocampus, middle cingulate cortex, PCC, precuneus, and calcarine sulcus, as well as the right posterior hippocampus and ACC, supplemental motor area, middle cingulate, superior and inferior parietal cortices, PCC, and precuneus (figure 2, supplementary table 2). We did not observe connectivity differences between first episode and chronically ill patients.

Fig. 1.

Areas of increased functional connectivity in unmedicated patients with schizophrenia compared to healthy controls (P _FDR < .05). Clusters were overlaid on the Xjview single subject T1 template, numbers indicate MNI coordinates. Color bar indicates t values. ACC: Anterior cingulate cortex; MPFC: Medial Prefrontal Cortex; MNI: Montreal Neurologic Institute.

Fig. 2.

Areas of decreased functional connectivity in unmedicated patients with schizophrenia compared to healthy controls (P _FDR < .05). Clusters were overlaid on the Xjview single subject T1 template, numbers indicate MNI coordinates. Color bar indicates t values. ACC: Anterior cingulate cortex; PCC: Posterior cingulate cortex.

Examining relationships between connectivity and clinical variables, we found that greater connectivity between the right anterior hippocampus and precuneus (t = 4.27, P _FDR < .01, x = −3, y = −54, z = 14) as well as lesser connectivity between the left posteriors hippocampus and precuneus (t = 4.14, P _FDR < .01, x = 12, y = −61, z = 17) and lingual gyurs (t = 3.73, P _FDR < .01, x = 18, y = −52, z = −4), was associated with greater negative symptom burden (supplementary figure 2). No relationship between connectivity and positive symptom burden was observed.

Functional Connectivity in SZ After 6 Weeks of Treatment.

Comparing connectivity patterns from every seed in SZ after 6 weeks of treatment to HC at week 6, we did not find and significant group differences in any of the seed regions.

Functional Connectivity in Relationship to Treatment Response.

Connectivity in Unmedicated Patients and Treatment Response

Connectivity between the left anterior hippocampus and left ACC and right caudate nucleus (t = 4.46, P _FDR < .01, x = −16, y = 33, z = 27; t = 6.71, P _FDR < .01, x = 21, y = 17, z = 15), and between the left posterior hippocampus and left auditory cortex (t = 5.77, P _FDR < .01, x = −49, y = −19, z = 15) in unmedicated SZ was positively correlated with treatment response (figure 3). Right anterior hippocampus and lingual gyrus connectivity in unmedicated SZ was negatively correlated with treatment response (t = 5.45, P _FDR < .01, x = 14, y = −49, z = −4; figure 4).

Fig. 3.

Functional connectivity in unmedicated patients with schizophrenia in relationship to clinical response after 6 weeks of treatment. Left column: Clusters indicate regions where lower connectivity strength with the hippocampus is correlated with good treatment response after 6 weeks (P _FDR < .01). Clusters were overlaid on the Xjview single subject T1 template, numbers indicate MNI coordinates. Color bar indicates t values. Right column: Correlations between baseline functional connectivity strength and change in positive symptom severity. ACC: Anterior cingulate cortex; BPRSpos: Change in BPRS positive subscale scores; L: left.

Fig. 4.

Functional connectivity in unmedicated patients with schizophrenia in relationship to clinical response after 6 weeks of treatment. Left column: Clusters indicate regions where lower connectivity strength with the hippocampus is correlated with good treatment response after 6 weeks (P _FDR < .01). Clusters were overlaid on the Xjview single subject T1 template, numbers indicate MNI coordinates. Color bar indicates t values. Right column: Correlations between baseline functional connectivity strength and change in positive symptom severity. BPRSpos: Change in BPRS positive subscale scores; R: right.

Connectivity Changes Over Time and Treatment Response

Examining hippocampal functional connectivity changes over time in relation to treatment response in the areas correlated with clinical response while unmedicated, we found both increased connectivity (supplementary figure 3; right posterior hippocampus with 2 clusters: MPFC/ ACC [t = 4.99, P _SVC < .05, x = −26, y = 32, z = 20], and caudate [t = 5.73, P _SVC < .05, x = 21, y = 16, z = 15]); (right anterior hippocampus and lingual gyrus [t = 4.71, P _SVC < .05, x = 16, y = −60, z = −12]); (right posterior hippocampus and lingual gyrus [t = 2.64, P _SVC < .05, x = 15, y = −50, z = −4]); and decreased connectivity (supplementary figure 4; right anterior hippocampus with 2 clusters: left auditory cortex [t = 4.77, P _SVC < .05, x = −51, y = −12, z = 12], and right caudate [t = 2.99, P _SVC < .05, x = −15, y = 2, z = 14]); (left anterior hippocampus with left auditory cortex [t = 4.69, P _SVC < .05, x = −56, y = −22, z = 14]; (left posterior hippocampus with left auditory cortex [t = 3.01, P _SVC < .05, x = −52, y = −14, z = 15]) were related to improvement of positive symptoms. Examining the association between connectivity change over time in relation to improvement of hallucinations, we found a similar pattern (supplementary figures 3 and 4).

Discussion

To better delineate hippocampal connectivity in schizophrenia and the effect of antipsychotic medication in this functionally heterogeneous area, we examined anterior and posterior hippocampus resting state networks in a longitudinal study design. Widespread aberrant connectivity with all 4 seeds, most prominently to medial cortical regions, was observed in unmedicated patients. We found baseline connectivity, specifically to the auditory cortex, lingual gyrus, caudate, and dorsal ACC, was related to clinical response after 6 weeks of treatment. Notably, connectivity patterns in these areas also changed over time as a function of treatment response.

Hippocampal Dysconnectivity

We previously reported abnormal resting state connectivity between the whole left hippocampus and the bilateral precuneus in a smaller, but, overlapping group of unmedicated patients with schizophrenia.¹⁷ With an expanded sample size and more detailed anatomical delineation, we again appreciated this abnormality, but also observed dysconnectivity with an extended midline region including the dorsal ACC, supplemental motor area, MPFC, middle cingulate, as well as the lingual gyrus, parietal cortex, and occipital cortex, similar to previous reports in medicated patients.^18,19 Notably, dysconnectivity was seen in both anterior and posterior networks, which is consistent with anatomic studies in schizophrenia reporting that volume loss is not regionally selective along the hippocampal longitudinal axis.⁴⁵

Resting State Connectivity and Treatment Response

As hypothesized, we found resting state patterns in unmedicated patients were associated with clinical response after 6 weeks of medication. Higher baseline hippocampal connectivity with the dorsal ACC, left caudate, and left auditory cortex as well as lower connectivity with the lingual gyrus were indicative of good response, consistent with our previous report that ventral tegmental area connectivity to the dorsal ACC and the default mode network in unmedicated patients with schizophrenia was predictive of the extent of symptom reduction with medications.³² It is striking that hippocampal connectivity with these areas has been associated with positive symptom burden in medicated patients.⁴⁶

Our findings also are consistent with reports linking abnormal activation in the hippocampus and auditory cortex with predisposition to spontaneous verbal auditory recollections or auditory hallucinations,^47,48 and a meta-analysis of cortical activations showing increased activity in speech perception and production areas as well as the hippocampus in those actively hallucinating.⁴⁹ Resting state connectivity strength between the left auditory cortex and left hippocampus was reported to be associated with the severity of hallucinations,⁴⁶ and a reduced coupling between the left auditory cortex and right hippocampus was found in those patients prone to auditory hallucinations.⁵⁰ In a study examining differential involvement of the hippocampal complex in visual hallucinations based on sensory modality, Amad and colleagues reported that patients with auditory and visual hallucinations showed increased connectivity of the hippocampus with the MPFC and caudate, when compared to those suffering from auditory hallucinations only.⁵¹

Our finding of a relationship between hippocampal-lingual connectivity and treatment response was not hypothesized a priori. Interestingly, an increased parahippocampal-lingual gyrification index in first episode schizophrenia has been reported to be inversely correlated with the age of onset of psychosis.⁵² Given that earlier age of onset is associated with poorer clinical outcomes,⁵³ it may not be surprising that we found hippocampal-lingual connectivity to be associated with treatment response.

Here, we observed changes in connectivity over a 6-week trial of antipsychotic therapy in aforementioned regions were correlated with good clinical response. Hippocampal connectivity reductions to the auditory cortex and increases to the lingual gyrus a function of successful treatment were observed. In contrast, greater hippocampal baseline connectivity followed by a connectivity increase over time to the dorsal ACC, MPFC, and caudate was associated with better response. Post hoc analyses suggested that findings were related to a reduction in hallucinations. A similar modulation of connectivity between the hippocampus and caudate as a function of clinical response, albeit without evidence of baseline dysconnectivity, has been reported by Sarpal et al³³ using striatal seeds. This discrepancy in baseline findings may be due to antipsychotic exposure in some of the patients prior to the initial scan in the Sarpal study. In schizophrenia, disruption of fronto-temporal connections has been flagged as a major dysfunction.^9,13,54–56 Extending these findings, we recently used Granger causality methods to evaluate effective connectivity between hippocampal and frontal seeds during a memory retrieval task.³⁴ We observed reduced effective connectivity from the hippocampus to the medial frontal cortex in unmedicated patients compared to HC, and increased effective connectivity from the hippocampus to all frontal regions after 1 week of treatment with risperidone. Interestingly, the network of regions associated with treatment response emerging in our studies recapitulates the circuit abnormality identified by Goto and Grace, who postulate that aberrant dopaminergic function, disrupting the balance between limbic and cortical drive, could result in delusions and cognitive dysfunction.⁵⁷

Hippocampal Pathology, Symptom Profiles, and Connectivity

Hippocampal structural abnormalities have been linked to functional markers, including different symptom dimensions, in both patients with schizophrenia^3,58,59 and those at ultra-high risk for the illness.^59–61 Here, we reported associations between hippocampal connectivity patterns and negative symptom severity as well as response to antipsychotic treatment. Increased hippocampal rCBF⁵ and glutamate levels³ might be a result of faulty GABA interneurons⁶² which are thought to generate oscillations in the gamma frequency ranges through which synchronized brain activity is regulated⁶³; their dysfunction could disturb synchronicity and affect resting state connectivity. Schobel and colleagues identified a relationship between hippocampal hypermetabolism and structural deficits in patients transitioning from a prodromal state to syndromal psychosis and concluded that glutamate acts as a pathogenic driver of hippocampal pathology,⁶⁴ but a direct link between glutamatergic/ GABAergic dysfunction and hippocampal resting state connectivity abnormalities in schizophrenia is yet to be established.

Mechanisms of Antipsychotic Drug Action

The link between treatment response and hippocampal resting state connectivity established in our study suggests possible mechanisms by which antipsychotics alleviate positive symptoms. Animal studies have shown that antipsychotics can be used to restore cortical synchronization and functional connectivity after disruption with hallucinogenic agents,^65,66 which is consistent with our previous findings of decreased hippocampal rCBF with antipsychotic treatment²⁹ and lack of glutamate elevation in medicated patients with schizophrenia as opposed to unmedicated patients.^3,30 Other mechanisms, such as changes in myelination of white matter tracts^67–69 or changes in neurotrophin expression⁷⁰ may also affect functional connectivity within neural networks.

Strengths and Limitations

Several strengths and limitations have to be considered in the interpretation of our findings. To minimize variance in the data, we carefully matched groups on several factors including parental socioeconomic status and smoking, enrolled only subjects free of exposure to antipsychotic medications for at least 2 weeks preceding the baseline scan, used a single antipsychotic medication, did rigorous preprocessing and motion “scrubbing” of scans,³⁹ and increased the signal to noise ratio by excluding datasets with EPI signal dropout in the seed region. Because we observed a significant change in connectivity longitudinally in 1 out of 4 seeds in HC, we only considered connectivity changes that were related to treatment response. Without a placebo group, it is impossible to definitively attribute connectivity changes to effects of treatment, but such a study design is not feasible, because known effective treatments cannot be withheld from patients.

Summary of Findings

In summary, we found widespread aberrant hippocampal connectivity along its longitudinal axis in unmedicated patients with schizophrenia and changes in connectivity with antipsychotic treatment related to reduction in positive symptoms. We speculate that excess hippocampal drive related to aberrant dopaminergic function may be associated with observed abnormalities, and implicate neural networks that could be targeted to evaluate potential new therapeutic interventions.

Supplementary Material

Supplementary material is available at http://schizophreniabulletin.oxfordjournals.org.

Funding

This work was supported by the National Institutes of Mental Health (NIH; R01MH081014 and R01MH102951) and the University of Alabama Health Services Foundation General Endowment Fund Scholar Award.