As described in the Sleep Disturbances section, polysomnography is regarded as the gold standard for sleep measurements. It comprises several measurements such as the electrophysiological measurement of brain activity, muscle and eye movement, as well as oxygen level and heart rate. This sleep measurement approach allows, for instance, for the detection of the transition from deeper sleep stages to lighter sleep stages or wake at any time during sleep. Moreover, polysomnographic analyses have the considerable advantage that they are not subject to biases due to, for example, personal attitudes and knowledge, as it is the case for self-rated sleep assessments.

^{Figure 1: Schematic representation of the electrode positioning for the measurement of polysomnography (source: DLR)
Figure 2: Preparing subjects for polysomnographic recordings (source: DLR)}

A combination of polysomnography with continuously recorded sound pressure levels at the subjects home allows an exact calculation of the effect of the noise of single aircraft noise events on sleep, in particular, awakening reactions due to it. Such controlled experimental studies provide a so-called event-related evaluation at any time during the night. Through this, it is possible to develop an optimal statistical model, focusing on all relevant acoustic parameters such as maximum levels, level rise speed, event duration and the background noise level when the aircraft noise event occurs, but also non-acoustic situational factors, such as the elapsed sleep duration, the stage of rather deep or relatively light sleep before the noise event occurs, the time spent in the sleep stage before the noise event as well as personal parameters (e.g. age). All this data is condensed in exposure-response models and curves derived from these models. Exposure-response curves deliver the probability of awakeness for every noise event depending on its maximum sound pressure level and control of the influence of the other acoustical and non-acoustical factors mentioned above.

This methodology is very costly and time-consuming.  Moreover, it requires that only subjects are examined without any inherent sleep disorders and without any diseases which have side effects on sleep. Otherwise, any sleep effects following an aircraft noise event and total sleep quality parameters cannot be interpreted unambiguously. Consequently, the examination of subjects only on the two- to lower three-digit number range can be performed, and only very few research groups worldwide investigate the effects of noise on sleep employing polysomnography and provide exposure-response functions. When the ANIMA standard prediction model for sleep disturbances due to nocturnal aircraft noise was developed, only two field studies were available using this polysomnographic measurement approach. Both studies were operated by the German Aerospace Center (DLR):

• The STRAIN (STudy on human Response on AIrcraft Noise) study that was conducted between 1999 and 2003 and examined the influence of nocturnal aircraft noise on sleep utilising field studies (64 residents in 9 successive nights each) around Cologne/Bonn Airport with a 24 h operating scheme;
   
• The NORAH study (NOise Related Annoyance, cognition and Health) was carried out between 2011 and 2015 around Frankfurt airport before and after the beginning of the operation of the fourth runway in October 2011 and the associated night flight ban between 23 h and 05 h.

Within the framework of ANIMA, a standard exposure-response model and derived curve were generated based on a re-analysis of these field studies. The ANIMA standard exposure-response model is more comprehensive and accurate due to the inclusion of additional acoustical parameters than in the original publications. The ANIMA re-analysis in both cases shows higher awakening probabilities than in the previous publications. Confidence intervals of the re-analysis are somewhat smaller due to the better-explained variance. The ANIMA exposure-response curve shows the association between the maximum sound pressure level and the probability for an aircraft noise-induced awakening reaction while taking into account also the effect of further acoustical and non-acoustical factors (Figure 3).

These additional factors comprise:

1. noise event duration;
2. level rise speed;
3. background noise before the noise event;
4. elapsed sleep duration;
5. sleep stage before the noise event;
6. age; and
7. an interaction term between the maximum and background noise levels describes how strongly the aircraft noise event emerges from the background.

^{_{Figure 3: Probability of an awakening due to an aircraft noise event as a function of the maximum sound pressure level indoors and the following additional predictors: a) noise event duration, b) level rise speed, c) background noise before the noise event, d) elapsed sleep duration, e) sleep stage before the noise event, f) age,  and g) an interaction term between the maximum level and the background noise level (Hauptvogel et al., 2021)}}

Since different types of airports were considered in the ANIMA-reanalysis, the standard prediction model is assumed to be generalisable over different types of airport and shall be used to calculate night noise protection concepts based on human sleep physiology. The model seizes on the protective assumption and, therefore, considers an 'awakening' not only as a transition to stage 'wake' but also as transition to the light and non-restorative sleep stage 1. Although not statistically significant, age also was included in the model since the sleeping structure changes with age, shifting to lighter and more fragile sleep in general.

Having information on the maximum noise levels and a number of night-time aircraft noise events available, the standard prediction model allows the estimation of the number of additional awakenings due to aircraft noise per night. This information on additional awakenings per night can be calculated for every household around an airport and can easily be communicated to both – the affected communities and stakeholders.

Based on the ANIMA prediction model for sleep disturbance due to nocturnal aircraft noise, a Virtual Community Tool software was created which allows to visualise and understand additional noise-induced awakenings around an airport area depending on different noise scenarios (e.g., different numbers and times of operations and different aircraft types).

^{Figure 4: Results showing the increased zone boundaries of one additional awakening for an increase of air traffic for hypothetic airport scenario}

However, the studies carried out to develop a night noise protection concept for aircraft noise and, thus, both – the prediction model and virtual community tool – have some limitations that should be kept in mind:

• The ANIMA prediction model is based on only two field studies. It is desirable to have additional sleep data sets from field studies applying physiological sleep measurements and simultaneous noise recordings at the residents' homes, allowing event-related analyses. Moreover, methodologies to measure noise effects on sleep at a large-scale level, including vulnerable groups such as elderly and sick people or children is needed;
• It is not completely clear how many additional noise-induced awakening reactions lead to increased health risks after long-term exposure to aircraft noise. It should be the first question for future epidemiological noise effects' research because the current studies do not have sufficient scientific evidence for this question.

Notwithstanding the above, existing results are sufficient to demonstrate that it is necessary to improve the residents' level of protection. It is time to shift the aircraft noise calculation system for the night from purely physical (annual) quantity indicators – that may be identical for both ten smaller aircraft and only a single large per night – to a measure that considers human physiology and every single event that can awake thousands of people in one night.