1.1.

Methods of Measuring Core Body Temperature

There is a definite correlation between body temperature and diseases.⁵ Many previous studies have analyzed core body temperature using rectal temperature although corneal, tympanic, or axillary temperatures may also indicate core body temperature. Even though rectal temperature is considered to be more accurate, the placement of a thermometer in the rectum is unpleasant and inconvenient. In humans, corneal temperature was found to be linearly related to body core temperature with corneal temperature apparently plateauing at 36.5°C to 37.0°C. The corneal temperature of a healthy person is generally between 32°C and 35°C, which is lower than the body temperature and higher than the atmospheric temperature. The axilla and tympanic membrane are the other easily accessible sites for temperature measurement.⁶ Among all the above mentioned methods, corneal temperature measurements in humans can be easily taken with relatively noninvasive procedures.

Continuous studies since 1970s have been performed to study the effect of serotonin on the thermoregulation. An overview of the evidence supporting a central role for serotoninergic neurons in the control of body temperature by the diencephalon was presented by Myers.⁷ Morphological analysis of the anterior hypothalamus manifested that hyperthermia was elicited in almost all species when 5-hydroxytryptophan was injected locally into this thermosensitive zone. Moreover, the heat production response during cold stress in animals such as rats and monkeys was disrupted following damage to the serotoninergic neurons in the anterior hypothalamic preoptic area.⁷ A study suggests that serotonin takes part in the control of the body temperature set-point.⁸ Some studies showed that efficacious treatment of MDD normalized the lowered amplitude of the temperature circadian rhythm and increased body temperature. For example, it was found that disturbed circadian temperature rhythm in major depression patients was restored by electroconvulsive therapy (ECT). The temperature profile over a period of 24 h was noticeably different in patients’ pre-ECT and post-ECT and in controls. The 24-h profiles exhibited by post-ECT patients were similar to healthy subjects. The amplitude of circadian temperature rhythm was elevated after ECT. The mean temperatures (both asleep and 24 hours) were substantially higher in pre-ECT than post-ECT and healthy subjects.⁹ Rausch¹⁰ presented a study which claims to be the first demonstration of higher daytime body temperature in cases with major depression. The study showed that subjects with a corrected temperature above 36.8°C were 2.6-fold more likely to be depressed.

1.2.

Body Temperature in Psychiatric Disorders

Since 1930, a number of studies have been conducted to provide insight into thermoregulatory disorders in people with SCZ. A review of the relevant articles regarding thermoregulatory dysfunction in SCZ patients was conducted.¹¹ The studies by Cameron¹² and Buck et al.¹³ found that SCZ patients have lower baseline temperature when compared to that of controls. The initial skin temperature recorded by Hermesh et al.¹⁴ was significantly lower in the patient group, but Douglass and Toogood¹⁵ found that the baseline temperature in SCZ patients and controls were similar, and the study of Shiloh et al.¹⁶ showed a higher baseline temperature in patients. A longitudinal study examining the thermoregulatory alterations induced by acute antipsychotic drug (APD) administration in SCZ patients showed that acute administration of APDs decreases body temperature in drug-free SCZ patients by $\sim 0.36°\mathrm{C}$, although the absolute values remain within the normal range.¹⁷

It has been reported that major depression is associated with increased core body temperature, reduced thermoregulatory cooling (e.g., sweating), and alterations in peripheral measures of serotonin (5-HT) activity, all of which are expected manifestations of impaired activity in the skin-to-brain-to-skin thermoregulatory circuit within which the ascending spinoparabrachial pathway and its CNS projections form core components.¹⁸ Furthermore, large effect size correlation was observed between reductions in Center for Epidemiologic Studies Depression Scale scores and reductions in mean core body temperatures.¹⁹

1.3.

Measurement of Corneal Temperature

Among the array of eye imaging systems, infrared (IR) imaging is relatively noninvasive, inexpensive, and harmless, and hence widely used in medical diagnosis. Thermography (or thermal imaging) can be defined as an imaging procedure which traces the thermal sequence with an IR camera. As a part of thermoregulation, human body emits heat to the atmosphere. As the body temperature rises, the amount of energy radiated at any wavelength increases and the wavelength of peak emission decreases. An increased temperature profile indicates underlying venous flow changes.⁵ The temperature distribution on the surface of the body can be acquired by thermographic camera to check the health condition of a person. As the temperature variation in the anterior segment of the eye rarely occurs, it has been used to study various eye problems.²⁰ However, emissivity may also change with the water content of the eye and of the cornea. The thermographic image allows one to see variations in temperature. The visual and quantitative documentation of the temperature measurements can then be made using the appropriate software.

1.4.

Corneal Temperature in Psychiatric Disorders

The corneal temperature of drug-free SCZ patients has been evaluated with a thermal imaging camera which demonstrated a significantly higher corneal temperature (1.55°C higher) in the patient group when compared to matched controls.²¹ Moreover, the study indicated that the corneal temperature has a positive correlation with the symptom severity assessed by the brief psychiatric rating scale (BPRS). Shortly afterward, a study showed that the corneal temperature of neuroleptic-treated patients was significantly lower than that of the drug-free patients.²² Another study then indicated that drug-free SCZ patients have significantly higher corneal temperature when compared to that of normal healthy controls or treated patients.²³ Based on these findings, Shiloh et al. conducted further studies to validate their hypothesis that the mental status of SCZ patients may be associated with their eye temperature. The stability of thermoregulatory variation in male treatment-resistant SCZ patients and healthy subjects was assessed over a period of 6 weeks;⁶ (to avoid changes in body temperature due to menstrual cycle, women were not included in the study). The corneal temperature was evaluated two to three times a week using a thermal imaging camera in a designated room. Shiloh et al. reported a significant correlation between SCZ patients’ mean weekly PANSS scores and their mean weekly corneal temperature.²⁴

Drugs may alter body temperature by acting on any component of the thermoregulatory system. These components include heat production, heat conservation, and heat loss effectors and their efferent pathways, thermosensors and their afferent pathways, and neurons within the CNS that coordinate thermoregulatory effector activities.²⁵

The current study aims at recording and investigating corneal temperature in patients with MDD and comparing them to healthy control subjects to study the relationship between corneal temperature and depression. We aimed to study the temperature profile across the cornea of the eye and to investigate the nature of corneal temperature difference in MDD patients and healthy controls. We also aimed to investigate if corneal temperature correlates to the degree of symptom severity.

2.1.

Participants

A total of 32 subjects aged 19 to 79 years [$\text{males mean}=46.88$ ($\mathrm{SD}=17.00$); $\text{females mean}=49.00$ ($\mathrm{SD}=12.05$)] comprising 16 subjects with MDD and 16 age- and sex-matched healthy controls underwent psychiatric and ocular thermographic assessments (Table 1). Subjects were recruited either through public notices or word of mouth from the clinical services of The Alfred Hospital, Victoria, Australia. MDD subjects were recruited before their transcranial magnetic stimulation (TMS) treatments began. Each was treated with either left-sided repetitive TMS (rTMS; 10 Hz) or right-sided slow (1 Hz) TMS. MDD subjects were required to have a diagnosis of major depressive disorder as determined by a treating psychiatrist and confirmed with the Mini-International Neuropsychiatric Interview (MINI)²⁶ and a score on the Hamilton Depression Rating Scale (HAMD)²⁷ of at least 13, representing moderate-to-severe depression. Exclusion criteria included another current or previous DSM-IV axis I diagnosis, current active medical problem, and known neurological disease. In addition, control subjects were required to have no history of psychiatric illness as screened by the psychiatric assessment schedule.²⁸ All subjects provided written informed consent on a form approved by the Alfred Hospital, La Trobe University, and Monash University Human Subjects Research and Ethics Committees.

Table 1

Demographics of the patient and control subjects.

Patients were taking a variety of medications, including SSRIs, SNRIs, MAOIs, and atypical antipsychotics.

No subjects had diagnosed eye disease (e.g., glaucoma and cataracts).

2.2.

Thermographic Camera

Because the normal core temperature of a human body ranges between 32°C and 38°C, the temperature measurement range required for this study is small. We used an NEC F30S Thermo Shot (NEC Corporation, Japan) which has a spatial resolution of $160\times 120\text{pixels}$, thermal resolution of 100 mK ($\pm 0.1°\mathrm{C}$), and accuracy of $\sim 2000\mathrm{mK}$ ($\pm 2°\mathrm{C}$). Emissivity was set to 0.98 based on previous corneal temperature studies.^29–31

2.3.

Room Temperature, Humidity and Lux Monitoring, and Corneal Thermographic Acquisition

All thermographic assessments were performed in a specified designated room with a temperature of 23.2°C to 26.4°C and relative humidity of 42% to 47%. Based on the specifications of our camera, the best measurement distance to image each cornea separately was evaluated to be 20 cm. We constructed apparatus to measure and record the temperature, humidity, and intensity of light at all times.

A National Instruments USB-6009 data acquisition device was used to acquire differential mode data from temperature (LM35; Texas Instruments), humidity (HIH 4000; Honeywell S&C), and light sensors (ORP12; Silonex, UK). LabVIEW (National Instruments, Texas), and an open standard graphical programming environment was used to access and control the data acquisition hardware.

Both eyes were imaged separately. Using a protractor, the images were taken at an angle of 15 deg between the eye and the camera to avoid variations in surface emissivity at the given observation angle. Moreover, the human eye has an ellipsoidal-type shape, so some variations and errors can occur in the IR image due to variation in the observation angle. By fixing the viewing angle, variations due to surface emissivity and observation angle were avoided. GORATEC Thermography Studio (GORATEC Technology, Erding, Germany) was used to provide an accurate detailed analysis of the images stored into the memory card. The analysis objects such as points, lines, and areas helped to create histograms and line profiles to present the thermal information contained by the images. To study the thermal information of the corneal surface, an elliptical area was used because it resembles the oval shape of eye. This area was selected manually by comparing with the RGB image provided by the camera and visual inspection of the temperature profile of the eye. The area of the ellipse (the physical imaging area) was fixed at 710 to $715{\mathrm{px}}^2$ ($\mathrm{35,400.42}{\mathrm{mm}}^2$) to ensure uniform analysis in all participants; regional areas were one quarter of the total area (i.e., quadrants). An area histogram showed the temperature distribution, indicating the minimum, maximum, average, and radiation power values across the cornea. The temperature profiles across the cornea were created using line objects in which the line profile (the length of which depends upon the area of the full eye) showed temperatures of every point along it. This indicated whether the temperature profile across the cornea of a depression patient was different from healthy controls.

2.4.

Procedure

Each participant participated in a single 1.5 h session which consisted of a cognitive and symptom rating session (including administration of the DSM-IV, MINI, and HDRS to patients with major depression disorder) of $\sim 45\min$, and a thermographic assessment of $\sim 25$ to 30 min. For the thermographic component of the study, participants were required to sit in a comfortable chair for $\sim 15\min$ with their chin stabilized in a chin-holder and eyes wide open. Participants were asked not to blink during the image acquisition periods. The thermographic camera (mounted on a tripod) then imaged various sections of each cornea to estimate their temperatures. The images were stored in JPEG format containing the radiometric data for analysis. The ambient temperature, humidity, and light intensity of the room were recorded so that they could be considered when calculating the ocular temperatures.

2.5.

Statistical Analysis

GORATEC Thermography Studio and LABView were used for detailed analysis and evaluation of thermal images. SPSS for Windows version 22.0 (SPSS, Illinois) was used to conduct statistical analyses between the groups. Analyses were two-tailed and were evaluated for significance at the 0.05 alpha level. Simple chi-square and analysis of variance (ANOVA) were employed to compare demographics and ocular temperatures, and Pearson correlations were used to investigate the association between clinical severity and ocular temperature.

3.1.

Demographics

There was no significant age or gender difference among the groups, hence combined male and female results are presented. Mean HAMD was 25.06 ($\mathrm{SD}=5.08$); two patients had scores below 20 (moderate depression), while the remaining 14 patients had scores beyond 20 (major depression).

3.2.

Corneal Temperature in Major Depressive Disorder and Controls

There were no significant differences in mean corneal temperatures between the MDD patients and control subjects [right cornea: $F(\mathrm{1,30})=0.624$, $p=0.436$; left cornea: ($F(\mathrm{1,30})=0.664$, $p=0.664$; Table 2]. There was no significant difference in temperature between the left eye and right eye of patients with MDD or in control subjects. Controlling for age made no difference to these results. Figure 1 shows an example of temperature measurement.

Table 2

Mean ocular temperatures of the patient and control groups.

Fig. 1

Example of thermographic software used to estimate corneal temperature.

3.3.

Corneal Temperature and Clinical Severity

There was no significant correlation between HAMD and age [$r(16)=0.87$, $p=0.748$]. All correlations between age and all ocular temperatures measurements were negative and none were statistically significant (all $p>0.05$). There were statistically significant positive correlations between right eye temperatures and clinical depression severity (Table 3 and Fig. 2). Controlling for age made very little difference to these results.

Table 3

Correlations between HAMD and ocular temperatures.

Fig. 2

Relationship between HAMD scores and ocular temperatures (whole eye averages).

3.4.

Corneal Temperature and Medication

Each patient was taking different medications, so statistical analysis of the potential effect of medication could not be clearly considered. However, two patients who were taking lithium carbonate had corneal temperatures that were among the lowest within the clinical group.

4.1.

Limitations

The study was limited by the number of participants, hence the results must be interpreted with caution. The study was also limited by the accuracy of the thermographic camera which may be insufficient for this study, noting that the temperature difference between MDD and normal groups may be beyond the thermal resolution of our camera. Therefore, the average corneal temperature of the patients with MDD may be found to be different from that of the healthy controls in more extensive studies recruiting higher numbers of subjects and using cameras of higher thermal resolution. Corneal temperature evaluation may then potentially serve as a noninvasive diagnostic tool to discriminate MDD patients from normal healthy people.

Measurements were not acquired at the same time of day for each subject, and ocular temperatures fluctuate during the day.⁴⁰ We conducted only one measurement per person, so we did not account for this diurnal variability. Furthermore, body temperature (and, therefore, ocular temperature) also varies with menstrual cycle.⁴¹ These factors may have impacted upon the reliability of our conclusions. Hence, we will endeavor to control for these variables in any future corneal temperature investigations.

As this was a cross-sectional study, we could not establish a direct cause–consequence relationship between corneal temperature and depression severity.