Light flashes and other sensory illusions perceived in space travel and on ground, including proton and heavy ion therapies

Abstract

Most of the astronauts experience visual illusions, apparent flashes of light (LF) in absence of light. The first reported observation of this phenomenon was in July 1969 by Buzz Aldrin, in the debriefing following the Apollo 11 mission. Several ground-based experiments in the 1970s tried to clarify the mechanisms behind these light flashes and to evaluate possible related risks. These works were supported by dedicated experiments in space on the following Apollo flights and in Low Earth Orbit (LEO). It was soon demonstrated that the LF could be caused by charged particles (present in the space radiation) traveling through the eye, and, possibly, some other visual cortical areas. In the 1990s the interest in these phenomena increased again and additional experiments in Low Earth Orbit and others ground-based were started. Recently patients undergoing proton and heavy ion therapy for eye or head and neck tumors have reported the perception of light flashes, opening a new channel to investigate these phenomena.

In this paper the many LF studies will be reviewed, presenting an historical and scientific perspective consistent with the combined set of observations, offering a single comprehensive summary aimed to provide further insights on these phenomena.

While the light flashes appear not to be a risk by themselves, they might provide information on the amount of radiation induced radicals in the astronauts’ eyes. Understanding their generation mechanisms might also support radiation countermeasures development. However, even given the substantial progress outlined in this paper, many questions related to their generation are still under debate, so additional studies are suggested. Finally, it is also conceivable that further LF investigations could provide evidence about the possible interaction of single particles in space with brain function, impacting with the crew ability to optimally perform a mission.

1. Introduction

The history of the investigations about light flashes (LF) perceived by the astronauts while in darkness covers 54 years of experiments and studies. It starts with the report by Aldrin about the LF he perceived during his first Moon mission, and reaches our days, with an interest in the worldwide scientific community varying with some synchronization to the human space exploration endeavors. Initially the research was motivated by the concern that LF could be the sign of possible risks affecting brain function. Today this concern might be over, and the interest is becoming mostly scientifically related, and aimed at developing radiation countermeasures. In the latest years proton- or heavy ion-therapy patients have reported seeing light flashes during therapy. This has started a renewed interest in this matter and offered a new channel to study these effects.

Visual perceptions occurring independently of physiological photonic stimulation of the retina are called phosphenes. This word describes a large number of symptoms related to different pathologies or brain statuses. We will refer to the phosphenes treated in this review simply as Light Flashes (LF), as this has been the name used to describe them since the beginning, and also to distinguish them from the other kinds of phosphenes.

There have been a few reviews in the past on this issue [1], [2], [3], [4], [5], the last one specifically on LF perceived during proton therapy. However, there has never been an attempt to discuss together the results of the different experiments (in space and on ground, based on an animal model, in vitro, during therapies) in order to evidence the existing and the missing knowledges in this area.

This review aims at fulfilling this gap, and presents a brief history of the works on LF, followed by a more detailed description of the studies connected with LF in space.

Finally, the discussion will try to describe the status of the knowledge today, and the relevant gaps.

2. A Light Flash short history

‘I observed what I thought were little flashes inside the cabin’ this short comment made by Buzz Aldrin on July 31st 1969 [6], started the LF interest. Even during the debriefing that sentence generated a debate. Armstrong, indeed, initially stated ‘I'd seen some light, but I just always attributed this to sunlight’ but shortly thereafter he also admitted ‘It could be something like Buzz suggested’. Aldrin made other remarkable observations in his debrief (issues that will be discussed later in this paper): ‘I was able to <omissis> see double flashes, at points separated by maybe a foot. At other times, I could see a line with no direction of motion’ and later in the debrief ‘a rather funny feeling to contemplate that something was zapping through the cabin’. With these remarks Aldrin anticipated several of the important questions, about LF motion and appearance.

The reason for these anomalous LF (anomalous in the sense they were perceived in the absence of ‘real’ external light) was soon hypothesized to be the cosmic radiation interacting somehow with the retina. This hypothesis was correct, and not unexpected: almost a decade before the first human spaceflight, in 1952, Cornelius Tobias studying the radiation field at high altitude stated that “It is conceivable that very densely ionizing tracks would produce small flash-like light sensations.” [7]. The LF reported by Aldrin is indeed the confirmation of Tobias hypothesis. Nevertheless, the variety of shapes and the movement reported by Aldrin were difficult to explain. Furthermore, the mere idea that cosmic radiation in space could alter visual sensations caused worries. If radiation in space could interact with brain function, how could we be sure that this interaction would just affect vision with possibly harmless flashes? Is there a possibility that this could be just the iceberg tip of an effect that could produce relevant health hazards for the astronauts? In the Biomedical results of Apollo, published in 1970 [8], the need of investigations to understand this phenomenon and its possible broadening with other effects were clearly stated (“Although the potential hazards to living systems from the heavy nuclei component of galactic cosmic radiation was recognized, very little active research was conducted until the crews of Apollo 11 and subsequent Apollo missions reported experiencing a visual light flash phenomenon”). Therefore, new experiments were performed on ground and in space in the last Apollo missions.

It should be stressed that the possibility of modulating brain functions by externally delivered energy has several demonstrations in the literature. As an example, Amassian and his colleagues [9] showed that the delivery of a magnetic field on the occipital area could interfere with the vision process making the subject temporarily blind. The question was if something like this, or even a riskier alteration, could really happen in a relevant manner with space radiation.

The strong NASA request for LF studies produced a large number of experiments in very few years, both on ground (see session 3) and in space (see session 4.1), most of them published in very high level journals. It was demonstrated that charged particles (even neutrons) could elicit LF, confirming the hypothesis that cosmic radiation was the cause of them. Conjectures were also provided regarding the responsible mechanisms and interactions.

During the Apollo missions the spacecraft was exposed to galactic cosmic radiation (see, for example, [10]), which comprises all ions from protons to heavy ions with energies from several eV to ≈10²⁰ eV, with a spectrum peaking at ≈1 GeV/n and an isotropic and continuous incidence.1 In Low Earth Orbit (LEO) the flux of GCR is reduced due to geomagnetic shielding. A high flux of trapped protons, with a directional incidence, is added during passages through the South Atlantic anomaly (SAA), a region over Brazil where the radiation belts reach lower altitudes. The interactions between this radiation and the atoms of the shielding materials (provided by the space vessel, the base, the suit, the human body itself) generate secondary radiation, like protons and neutrons, and target and projectile fragments. Solar Particle Events (SPE), another important component of space radiation, when large fluxes of energetic protons are released by the sun, never occurred during all LF experiments.

Until today the LF generation mechanism issue has not been answered with the details it deserves scientifically. Now we are more comfortable due to a sort of epidemiologic evidence: with so many astronauts who flew in space for different periods, some of these for as long as one year, and with no evidence of LF related pathologies, the LF question appears to be pushed back in the scientific domain and not anymore of concern for space operations. Is this a correct attitude? This paper would also like to provide support to answer this question.

With the end of the Apollo program (December 1972) the interest in these effects dropped. After the first two investigations in LEO (see session 4.2), Skylab in 1973–1974, and Apollo Soyuz Test Program (ASTP) in 1975, the interest went to zero and it remained dormient till 1995, well into the MIR station era (1986–2000) when the investigations in space started again on the MIR, involving 6 astronauts (session 4.3). The studies moved then to the International Space Station (ISS) where 4 astronauts participated to different experiments (session 4.4). A summary of all the LEO experiments can be found in session 4.5. In parallel, at the beginning of this century, new ground based studies started (session 5).

At the same time a new powerful cancer therapy approach, hadron-therapy (including proton-therapy), was fast becoming an important therapy tool around the world. Patients started to report the perception of LF during therapies involving tumors close to the eyes and so studies on LF began to involve them (session 6).

Therefore, the panorama of the interests in LF slowly shifted from ‘only space’, also to hadron-therapy. In the therapy context other sensory illusions (not only visual) were reported by the patients. An increasing amount of publications about LF in the latest years regarding proton irradiations for cancer therapy were released.

3. The ground experiments in the 70s

Following the Apollo 11 LF report, controlled experiments on ground were carried on. The strategy was to use particle accelerators, position the experimenter’s eye in the beam line and record his reports about the possible LF sensations. The particle and energy were mostly defined by what was available in the accelerators active at the time (beginning of 1970). The other parameters that could be chosen by the researchers were the target (almost always the eye), the direction and the rate (kept as low as possible, for safety reasons and to best mimic space radiation). In Table 1 a summary of these results extracted from a former review [4] is presented. For more details the reader is addressed to the original review or to the original papers.

Table 1.

Observations of LF in ground experiments in the 1970s (Extracted from [4]).

Ref	particle	Energy / Momentum	Number / Fluence	Beam direction	LF
[14]	Cosmic µ	n.a.	single	≈F	Y
[15]	Cosmic µ	n.a.	single	F B L	Y
				L(cort)	N
[16]	N	3 MeV	105	F B	Y
[17]	N	3 MeV	2 × 104	F L	N
	N	14 MeV		F L	Y
				B	Y
[18]	n	20–640 MeV	1.4 × 104	F L	Y
	π+	1.5 GeV/c	200		N
[19]	He	5.3 MeV		(bulloks’ eye)	Y*
[20]	µ	6 GeV/c	3 × 103/pulse	L	Y
				B	N
[21]	n	< 25 MeV (pkd at 8)	103–104	F	Y
				B	Y
[22]	He	240 MeV	10–20/pulse	L	Y
	N	266 MeV/n	900	Cortical	N
[23]	N	531 MeV/n	≈1 / pulse	25° from optic axis	Y
[12]	µ	7.2 Gev/c	1–400/burst	F	Y
	π−	725 MeV/c
[24]	µ	7.2 GeV/c		F L B	Y
[25]	Π	725 Mev/c	1–1000/burst	F	Y
[13]	C	470 MeV/n 595 Mev/n	1–2 /burst	60° from optic axis	Y

*scintillation detected by photomultipliers.

F: front; L: lateral; B: back.

It was already known that one of the causes of LF could have been the Cerenkov effect: ions entering in the media (vitreous of the eye) with a speed higher than that of light in the same media are decelerated and emit radiation that can be in the visible spectrum. In this case the LF would be partly physiological as the visual system would perceive them as real visible photons. The only anomaly would be that the photons are generated inside the eye. Cerenkov effect is also possible with photons, and reports of LF from X rays Cerenkov radiation can be found even in nineteenth century (see references in [11]). In most Cerenkov related LF studies the LF were described as diffuse bluish light. To produce a visible photon an ion must have an energy higher than about 0.5 GeV/n. This value is easily found in cosmic rays, that have spectra peaking at ≈1 GeV/n. However, the characteristic bluish diffuse aspect of Cerenkov phosphenes does not match large part of the astronauts’ reports, starting with Aldrin’s debrief. Several of these ground experiments at the beginning of the 70s have been designed to study the possible incidence of the Cerenkov LF, some have been conducted at energies above and below the Cerenkov threshold [12], [13]. These studies confirmed that LF can be elicited at energies well below this threshold, with the characteristics reported by the astronauts and that therefore Cerenkov was probably just one of the many production mechanisms for LF.

These accelerator works involved muons, pions, neutrons and ions (helium, carbon, nitrogen), input energies from few MeV/n to several GeV/n and number of particles per burst going from single ions to 10⁴ ions. Frontal, lateral and back entrances in the eyes/retina have been used [12], [13], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25]. In one instance a trajectory going into the visual cortex (nitrogen, 266 MeV/n, ≈1 kHz) was used, with no production of LF [22]. Finally, a couple of experiments [14], [15] showed that cosmic muons on ground could elicit LF too, even though this could not be confirmed later [18]. LF elicited by bursts of relativistic muons have been reported in an experiment at Brookhaven AGS accelerator [30] showing that also small values of LET (Linear Energy Transfer, the energy released in the tissue) could be enough to produce LF.

This wide multidimensional parameter space (particles, energies, targets, etc.) has been explored very sparsely during the many experiments on ground, producing results and suggestions that a few times appeared in contrast with each other. Furthermore, about half of the ground studies used radiation with limited space relevance (muons, pions, photons). At that time the only accepted explanation for the appearance of LF was Cerenkov radiation, therefore several studies were mostly devoted to Cerenkov related effects. Nevertheless, considerations summed up with a confirmation of cosmic radiation as the reason for these anomalous perceptions, with the eye as the most likely target. Although the characteristic diffused bluish shape of the Cerenkov induced LF was rarely reported in space, Cerenkov radiation was still considered the cause of a minority for LF. Finally, in many papers there was a most often implicit, but sometime also explicit, suggestion for the need of heavy ions to elicit LF. However, as mentioned, the ground experiments have been limited on the use of only three ion species, with no test with protons and, furthermore, relativistic, very light particles such as muons were demonstrated to produce LF.

In a later paper, olfactory sensations caused by radiation were reported [26], establishing the first evidence that some cortical mechanism, beside the ones involving just the eyes, could be needed to explain these illusions, and should be looked for.

In the same years, new experiments were conducted in space, on the following Apollo flights, as well as on the Skylab and during the Apollo Soyuz Test Program (ASTP).

4. Light Flashes in space

After the first report by Aldrin (Apollo 11, 1969), several experiments have been conducted in the following Apollo flights (1969–1972), in the Skylab (1974) and in the Apollo module during the Apollo Soyuz Test Program (ASTP, 1975). The interest in the LF phenomena raised again during the MIR Station life (1986–2000) and experiments resumed in 1995, in the MIR station, to end in 2007 in the International Space Station.

Fig. 1 shows a panorama of all the manned (orbital, lunar) flights, the solar cycle dynamic during the flights and the time intervals when the LF measurement in space were conducted.

Figure 1.

From the bottom: number of manned orbital (or lunar) flights (USA, CCCP/Russia), sunspot numbers (describing the solar activity, and therefore inversely linked to GCR flux) and relative solar cycle number. The time when measurements in space of LF are performed is also indicated. Major accidents are indicated in red.

We will describe in this section these flight experiments.

4.1. The LF measurements during Apollo flights

During Apollo 14 15 16 17 flights LF counting experiments were carried on. Furthermore, a rather ingenuous hardware (ALFMED, Apollo Light Flash Moving Emulsion Detector) was sent in Apollo 16 and 17 flights in the first attempt to directly measure the particle(s) giving rise to the LF phenomenon.

The results from the LF experiments in the Apollo 14-17 missions are collected in one paper [27] while a previous paper [28] made numerical considerations about LF experiences reported during the Apollo 11 12 13 debriefing, claiming that the Cerenkov mechanism “may alone be responsible for the phenomenon” and proposing a single possible alternative: direct excitation of the photoreceptors by ionization. As mentioned above, studies performed a few years later showed that Cerenkov is indeed a possibility but probably explains only a very small fraction of the perceived LF.

In the abstract of the Pinsky paper [27] we read “The crew members on the last seven Apollo flights observed light flashes that are tentatively attributed to cosmic ray nuclei (atomic number ≥ 6) penetrating the head and eyes of the observer”. In the Biomedical Results of Apollo, the sentence is even stronger “It is believed that these light flashes result from high energy, heavy cosmic rays penetrating the Command Module structure and the crewmembers’ eyes”. From this tentative suggestion this idea propagated in the scientific community. This resulted initially in an effort to study the effects mainly of HZE (High Z, High Energy) particles, further confirming, at least apparently, the HZE need to produce LF.

During the Apollo flights, each observer performed individual 1-h sessions. The flashes were mostly described as ‘white’ or ‘colorless’, with one exception (Apollo 14 Lunar Module pilot Edgar Mitchel) who reported a ‘blue with a white cast’ flash.

‘Spot’ or ‘starlike’ were the most prevalent shapes of the perceived LF, followed by ‘streak’ (line) and then by ‘cloud’ (perceived always in the periphery of the field of view). There was no prevalence of one eye over the other and time wise these LF were randomly distributed. The interval between LF followed the Poisson distribution [29], [30]. In this last work an interesting study merging information from ground investigations and measurements during the Apollo flights can be found.

During the Apollo flights there has been a careful investigation on the needed dark adaptation time before being able to perceive LF. Operatively this was defined by the time a blindfolded subject took to perceive the first LF (see Table 2). This time was roughly stable around 10 min for Trans Lunar Coast (TLC) and 20 min in trans Earth Coast (TEC), probably suggesting that the LF in TEC being of less intensity (as it was also reported) they required a longer dark adaptation time to be perceived.

Table 2.

The results of the LF experiments during the Apollo flights. The three (two for Apollo 17) measurements for each mission are performed by the three astronauts in the crew. The rate (LF/h) is calculated as the average across astronauts, the error is the standard deviation, and provides a suggestion for the interindividual variability of this effect. ‘Dark’ columns indicated the time of first LF perception, assumed to be the needed dark adaptation time for that subject. TLC: Trans Lunar Coast; TEC: Trans Earth Coast. Extracted from [27]. For more information the reader is addressed to the original paper.

		TLC				TEC
Apollo mission	Length (min)	Dark (min)	LF	LF/h	err	Dark (min)	LF	LF/h	err
14	47					29	12	20.4	5.5
	47					17	22
	47					18	14
15	60	10		19.7	5.5	30	8	7.7	1.2
	60	9				26	9
	60	10				17	6
16	60	0	47	34.5	12.5	0	21	9.7	8.7
	60	0	22			21	8
	60						0
17	60	15	17	14.0	3.0		0	0	0
	60	39	11				0

The overall averaged LF rate after dark adaptation was 2.9 min (corresponding to 21 LF/h), and this was said to be “compatible with the hypothesis that cosmic ray nuclei of Z≥6 cause the light flashes”.

Sessions during TLC appear to have a reported LF rate larger than that perceived during TEC, as measured by the shorter mean interval between consecutive perceived LF (see Table 2 and Fig. 2). Apollo 17 measurements during TEC resulted in no LF at all for all crew members during the 1-h session (even though they did see some LF before and after). It is interesting to read the words of the lunar module pilot Harrison Schmitt, who claimed during the technical debriefing: “Transearth <omissis> We had light flashes just about continuously during the whole flight when we were dark adapted. I had one which I thought was a flash on the lunar surface. That one period of time when we had the blindfolds on for the ALFMED experiment there were just no visible flashes, although that evening, that night, before I went to sleep I noticed that I was seeing the light flashes again. So, it just seemed to be that one interval either side of it where the light flash was not visible to myself or to the other two crewmen.” This outlines the impressive intraindividual variability of these phenomena, that must be taken into account when interpreting this kind of data.

Figure 2.

LF perception rate during Trans Lunar Coast (TLC) and during trans Earth Coast (TEC) (see Table 2). The rate (LF/h) is calculated as the average across astronauts, the error is the standard deviation, and provides a suggestion for the interindividual variability of this effect.

The authors checked all the physical possibilities for different GCR flux in the cabin during the TLC and TEC periods and concluded that no physics explanation for this phenomenon could be provided.2

The measurements performed with ALFMED are briefly mentioned in the paper, as ‘analysis in progress’. Some results can be found in a NASA report [8]. ALFMED is a helmet-like device supporting GCR sensitive emulsion plates around the head. Two sets of plates were used. The first fixed with respect to the head. The second that could be set in slow translational motion (10 µm/s) during the measurement. During the 1-h session the plate would have travelled for 3.6 cm. The particles which produced aligned tracks in the two plates (the fixed and the moving ones) were discarded as not candidates because they passed before the plate was set in motion. The translation distance of the two hits by the particle was used to provide the time of occurrence of the event. Only particles that were compatible with a trajectory passing through the eyes were studied. Accurate analysis of the images on the photographic plates was said to be possible to get a good evaluation of the Z value and, in certain energy range, of the kinetic energy of the ion. Two candidates for ‘LF generator’ were found and both appeared to have Z>8. Several difficulties in interpreting the data are mentioned in the Biomedical results of Apollo. These results have never been published in peer reviewed journals.

At the end of the Apollo flights the widespread opinion was that the most likely explanation for LF were a sort of ‘direct neural stimulation’ in the retina. Cerenkov effects was surely giving rise some of the LF perception, but could not be the major contribution.

4.2. First LF measurements in Low Earth Orbit (LEO)

Immediately following the last Apollo mission, two important space LF experiments were performed: in the Skylab, 1974 [31], [32] and during the Apollo Soyuz Test program (ASTP), 1975, [33].

Both LF experiments used radiation detectors to monitor the environmental radiation. The used detectors were already on board (Skylab experiment) or dedicated, (silver chloride, cadmium doped crystals and a silicon telescope) for the ASTP experiment. These were the very first investigations on LF in LEO, where the Earth magnetic field heavily modulates the flux of GCR and is responsible for the generation of the radiation belts. The radiation trapped in these belts are mostly protons and electrons. A region of particular importance in LEO is the South Atlantic Anomaly (SAA), approximately above Brazil, where the radiation belt is much closer to the Earth surface reaching few hundred kilometers from ground. Spacecrafts in LEO cross this region up to four times a day.

The modulation of the Skylab LF rates with the geomagnetic cut-off (and therefore with the GCR flux) appears evident in the Hoffman report [32], supporting the idea that GCR cause the LF.

In the Skylab measurements (see Table 3), 10 min at the start of each session were allocated as dark adaption (this time is not considered in Table 3). In the first Skylab session no particular head position was specified, whereas in the second session the anterior-posterior axis of the head was aligned with Earth magnetic field lines in the SAA, in the attempt to maximize the effect of trapped radiation on the optical structures.3

Table 3.

Numbers of LF observed in experiments on Skylab 4 and on the ASTP. The time in the SAA is calculated extracting the values from [31]Fig. 1 and [32] Figs. 14.2, 14.3, [33] Fig. 12.6, 12.8. The crossing in the SAA was on the edge of the region in the first session of the Skylab 4 mission, and on the SAA center (with a flux higher of about one order of magnitude) in the second session. The ‘time no SAA’ columns are calculated from the total time, subtracting the SAA time and the dark adaptation time (10 min for Skylab). For all these reasons the LF/h in the SAA, italic values (but also outside the SAA) should be taken with great caution.

experiment	Time	Altitude (km)	Time no SAA (min)	LF no SAA	LF/h no SAA	Time SAA (min)	LF SAA	LF/h SAA
Skylab 4	1974 Jan 28	443	52	12	13.6	8	12	90
	1974 Feb 04	443	33	32	50.5	12	112	560
	Total		85	44	31.1	20	124	372
ASTP	1975 July	225	174*	79	27.2	6	3	30

* two astronauts performed the experiment in parallel for 87 min (outside the SAA) and the number of LF is referred to both astronauts. So the LF/h is calculated using double time (87 × 2 = 174 min).

It is possible to appreciate [31], [32] that during the first session only the edge of the SAA was traversed, while in the second session the orbit crossed its center where the flux is highest. The ‘time SAA’ column for the Skylab sessions is evaluated from these papers ([31] Fig. 1 and [32] Figs. 14.2, 14.3). The ‘Time no SAA’ is calculated by the time given in the papers, subtracting the dark adaptation time and the SAA time. For what said the LF/h should be taken with great caution, especially in the SAA, italic values in the table, where a possible error in time, combined with a large difference in the SAA flux between the two sessions, would produce a significant change in LF rate.

The remarkable difference between the two sessions in the Skylab experiment outside the SAA are difficult to explain as the cosmic ray flux is about the same for both sessions (and the astronaut is the same). In this case intraindividual variability may play a role. With the caution indicated in the previous paragraph, the higher number of flashes in the SAA during the second session could be explained because of a larger proton flux (about one order of magnitude), of the longer time over the SAA, and because of the care taken in the alignment of the head with the magnetic field lines. This again is supporting the idea of a particle-driven effect, even if, in this case, protons would play the key role while the most accepted idea at the time was that HZE ions were needed.

The experiment performed during the Apollo Soyuz test Program (ASTP) in 1975, in the Apollo module, was performed by two astronauts in parallel (in Table 3 the combination of their findings). The time spent in the full experiment can be inferred from [33] Fig. 12.6 (93 min) and similarly from Fig. 12.8 the time spent over the SAA (6 min). The same caution mentioned above should be taken for the rate values.

From the orbital distribution in Fig. 12.5 [33] the modulation with the geomagnetic cut-off appears weaker than in the Skylab, with higher rates closer to the south pole. However, the average LF rate outside the SAA is compatible with the Skylab one averaged in the two sessions, even if ASTP was flying at a lower orbit and the Apollo module was more shielded than the Skylab. The sharp decrease of LF in the SAA (no LF reported) could be partly explained with these two causes. Likely, also the directionality of the SAA trapped radiation and the relative head alignment during the second Skylab session, should be considered.

Apollo missions, LEO experiments and ground-based investigations were clearly confirming that LF were a consequence of cosmic radiation interactions with the visual system. Eye and retina were thought to be the sites of this interactions, even if, sporadically, other sites (optic nerve, visual cortex) were mentioned, however with no experimental evidence. Long term health issues seemed to be ruled out.

From Fig. 1 the long USA interruption of the space flights after the ASTP can be appreciated. Also, the interest in the LF phenomenon decreased, and this reduced attention was also supported by the consolidated idea that there was actually no real risk for the mission linked with these anomalous perceptions. Soviet missions, instead, kept going at an almost unaltered rate. And it has been from them that the quest to understand the LF phenomenon gained new momentum. The new Space Station MIR seemed to be the place to conduct further experiments on the issue.

4.3. Renewed interest on the MIR

The experiments on the MIR mostly aimed at improving the ALFMED experiment in the Apollo capsule using new technologies. The strategy was to use advanced active silicon detectors to measure and track particles passing through the eyes of the astronauts while recording the instant when LF were perceived. The detectors, SilEye, positioned in the Core module of the MIR, and later SilEye2, operating in the Kristal and Priroda moduli, have been acquiring data on the MIR Space Station between 1995 and 1999. By duration it is still the most extensive study of the LF phenomenon. Avdeev and coworkers [34] claim that one of the major goals was to “investigate the hypothesis that heavy ions are the dominant source for LF “. The final conclusive statement was “It has been shown that nuclei and largely ionizing particles are the dominant sources of light flashes in space”. Noticeable, the “heavy” adjective was dropped.

Both detectors were 6 planes (6 × 6 × 0.038 cm³) striped (16 strips, 3.6 mm wide, each plane alternately in the X and Y directions) silicon telescopes detecting passing through particles, able to measure the energy released in each silicon plane (LET, Linear energy Transfer) and to track the particles. For SilEye-2 two 1 mm iron absorber were placed between couples of X-Y silicon planes.

The detectors were positioned close to the head to capture a portion of the radiation field passing through the eye(s) of the astronaut, with the aim to possibly correlate a specific ion to a LF perception.

LF observation sessions involved a total of 6 astronauts for more than 25 h of measurements (Table 4). Lower LF rate is measured in the later SilEye-2 experiment.

Table 4.

SilEye LF experiments on MIR station (extracted from [34]).

Years	Altitude (km)	Astronauts involved	sessions	Total time (min)	LF	LF/h	detector
1995/96	400–415	3	9	492	87	10.6	SilEye
1998/99	400–355	4	17	800	116	8.7	SilEye-2
1998/99	400–355		3	250	30	7.2	No detector
TOTAL		6	26	1542	233	9.1

Interestingly, the descriptions of the LF by four of the participating astronauts rule out the possibility of Cerenkov radiation as generation mechanism (as mentioned the Cerenkov radiation is mostly diffused, bluish). On the contrary, the reported types were i) continuous line, ii) line with gaps, iii) shapeless spot, iv) spot with bright nucleus, v) concentric circles. With the first two making up ≈90% of the LF. This supports the reports during previous flights, in the 70s.

One of the astronauts took part on both SilEye and SilEye-2 measurements and reported a lower LF rate in the SilEye-2 experiments (from 9.7 to 7.6 LF/h, showing that the decrease in rate in Table 4 is probably not due to interindividual differences). Several reports in these SilEye sessions seemed to suggest a decrease in LF sensitivity during “the first one or two weeks in space”.

The rate of LF perceived during the passages in the SAA is reported to be “only slightly higher than the rate outside”.

Careful analysis of the particle data produced the estimates of “very strong candidate of nuclei as initiators of LF” indicated in Table 5.

Table 5.

SilEye-2 candidates as LF initiators.

Z	# of occurrence
2	3
2–3	2
8	2
24	1

A successive analysis of the same data, claiming that “frequencies of light flashes within the SAA are consistently higher than those for equivalent particle rates outside it” suggests a further contribution for LF inside the SAA. The authors of this paper provide a possible explanation: “direct interaction of heavy nuclei with the retina, causing ionization or excitation, and proton-induced nuclear interactions in the eye (with a lower interaction probability) producing knock-on particles” [35].

4.4. On the International Space Station

ALTEA (Anomalous Long Term Effects on Astronauts) has been the first and still unique program to systematically study the LF phenomenon [36], [37], [4]. It featured a set of ground experiments together with a novel space detector system to measure particles impinging in the brain with a large brain covering factor (up to 15–20%), to be used in the ISS, concurrently with an ElectroEncephaloGraph (EEG) to record the electrophysiological responses of the astronauts, as a possible objective determination of the LF.

Alteino [37], [38], an ALTEA precursor, featured the first 16-channel EEG to master the use of these measurements in space. Alteino was a further upgrade of the SilEye silicon telescopes and was developed for the ISS as a first step of the ALTEA project. The detector (sometime also called SilEye-3) was built similarly to SilEye-2, without the iron absorber and with 4 double (X and Y) planes instead of 3. Two scintillator planes sandwiching the 8 silicon planes provided enhanced triggering capabilities.

During a LF session, the astronaut positioned the head in the proximity of the detector, after having donned the EEG cap. Eight LF session were conducted by a single astronaut obtaining an average of 5.7 LF/h (see Table 6). No significant results from the EEG were reported.

Table 6.

Results of the LF experiment with Alteino on the ISS) data from [37]).

Year	Altitude (km)	Astronauts involved	sessions	Total time (min)	LF	LF/h
2002	390	1	8	461	44	5.7

Finally, also during the Alteino LF experiment “No significant increase of rate was observed while passing through the South Atlantic Anomaly” [37]

The ALTEA program also included a collection of subjective reports: a 29-item questionnaire was provided to 98 astronauts [39]. 59 of them, representing Shuttle, MIR and ISS crews, reported their answers which have been studied. Some of the most important results are listed below (multiple answers were allowed):

• –47 perceived LF
• –41 perceived LF in darkness, 11 in dim light and 2 also in bright light
• –there was no workload correlation
• –39 perceived LF in motion
• –There was never perception of vertical motion
• –Many perceptions were in color: 47 (white) 6 (yellow) 3 (orange) 3 (blue) 2 (green) 1 (red)
• –8 perceived an increase in rate and or intensity over the SAA

With the appropriate caution due to the retrospective nature of this investigation, a few important points were raised for the first time: the astronauts never perceived vertical motion, some did perceive them even in bright light, and all colors appeared to be present in the “LF-color palette” even if with a strong predominance of white LF (see also Fig. 7 in session 6.2).

Figure 7.

Single and multiple illusions perceived by the 19 patients during the study at Loma Linda.

The ALTEA detector system [40], [41] was composed by six identical silicon telescopes (named Silicon Detector Units, SDUs) each similar to the Sileye’s. Each SDU was a 6 striped (alternatively in X and Y directions) planes self-triggering telescope. The six telescopes were initially positioned in a helmet holder to cover with them a large portion of the head (15–20%). The self-trigger allowed to detect and measure ions releasing 3–800 keV/µm in silicon (corresponding to 25–45 MeV protons, 25–250 MeV/n helium ions and all the passing through heavier ions, up to relativistic Molybdenum). A 32 channel EEG was used to measure the electrophysiological activity. The EEG cap was worn before sliding into the 6 detectors helmet (Fig. 3). A three-button pushbutton was used to record the instants of the LF perception in a common time frame with the EEG and the particle detector. A visual goggle was worn to provide dark adaptation and to perform standard visual stimulation (the same stimulation was performed, as baseline, on ground) to test the behavior of the visual system in orbit [4]. LF experiments were conducted in the USLab of the ISS in late 2006 and early 2007. These have been the last LF measurements in space.

Figure 3.

An astronaut performing LF measurements with the ALTEA system (From Narici 2008).

Three astronauts performed a total of 7 sessions (Table 7, Table 8 most of them 1.5 h long, one orbit, with the first 0.5 h devoted to the visual stimulation session). One astronaut session was about 2 h in order to pass through the SAA (none of the other sessions did).

Table 7.

ALTEA LF measurements in the ISS (from [4]).

Years	Altitude (km)	Astronauts involved	sessions	Total time (min)	LF	LF/h
2006–2007	400	3	7	414	20	2.90

Table 8.

ALTEA: the differences across astronauts. The large variability weakens the meaning of the average across the astronauts shown in Table 7, the session duration does not include the ≈30 min devoted to visual stimulation (from [4]).

Astronaut	Session duration (min)	LF	LF/h
1	58	12	12.4
2	65	4	1.30
2	63	0
2	62	1
2	87	1
3	18	1	1.52
3	61	1

An analysis of possible ion candidates as LF initiators has been also performed with the ALTEA detector. Three candidates have been found, all with Z either 3 or 4.

Electroretinogram traces of these candidates showed a morphology quite like visual evoked potentials [4]. Recently, a new advanced data analysis of these electroencephalographic ALTEA data has been started, aimed at studying possible particle modulation of the astronaut electrophysiology also unrelated with LF [42].

ALTEA was the first to report a directly measured ion-flux through the eyes of an astronaut [4], and this is a significant data point as all LF generating models rely on the ions passing through the eye/retina to predict LF rate. The number of ions through both eyes (each modeled as a 2.5 cm radius sphere) were measured (in 2006-7) to be ≈18 per minute. However, the ALTEA narrow protons and Helium ion energy acceptance window, must be taken into account. This became possible with a later work [43] where a corrective factor to apply to the ALTEA measurement has been calculated through a detailed cross calibration with the detector DOSTEL. This leads to ≈24 ion/s passing in the two eyes. This result can be compared with the calculation from the flux measured by the detector LIDAL in 2020-22 [44]. Taking in consideration the average across the three different directions inside the ISS, it can be calculated 33 ± 7 ions/s in both eyes (again each modeled as 2.5 cm radius spheres). The agreement is good considering that the first measurement was performed close to a solar maximum and the second closer to a solar minimum.

4.5. A LEO summary

A summary of the LEO measurements is provided in the Table 9, Table 10 and plotted in Fig. 4. Table 10 presents part of the data considered, in the average, in Table 9.

Table 9.

LEO LF measurements, averaged for each mission/detector (see Table 3, Table 4, Table 6, Table 7).

Years	Altitude (km)	Astronauts involved	sessions	Total time (min)	LF	LF/h	Position Mission/detector
191974	443	1	2	85	44	31.1	Skylab 4
1975	225	2	1	174*	79	27.2	ASTP (Apollo capsule)
1996	400–415	3	9	492	87	10.6	MIR Core Module SilEye
1999	400–355	4	20	1050	146	8.34	MIR Kristal & Priroda SilEye-2
2002	390	1	8	461	44	5.7	ISS Pirs Alteino
2007	400	3	7	414	20	2.90	ISS USLab ALTEA

Table 10.

Measurements extracted from the average presented in Table 9: 1996–1999, results from the same astronaut who participated to both campaigns on the MIR station. 2006–2007 single astronaut measurements in the ISS, ALTEA experiment.

Notes	Years	Altitude (km)	sessions	LF/h	Position Mission/detector
The same Astronaut in SilEye and SilEye-2 (green in Fig. 3)	1996	400–415	1	9.7	MIR Core Module SilEye
	1999	400–355	1	7.6	MIR Kristal & Priroda SilEye-2
The three astronauts involved in ALTEA (red in Fig. 3)	2006	400	1	12.4	ISS USLab ALTEA
	2007	400	4	1.30	ISS USLab ALTEA
	2007	400	2	1.52	ISS USLab ALTEA

Figure 4.

number of LF per hour measured in LEO in the different missions. In blue the average for each mission. In green the measurements by the same astronaut during SilEye and SilEye-2 experiments (MIR, respectively in the Core and Kristal/Priroda moduli). In red the individual LF rate measured by the three astronauts involved in the ALTEA measurement in the ISS (USLab modulus). The higher red point is the single astronaut perceiving 12 LF during his session. The error bars (blue and red) are just counting errors and therefore represent a lower error estimate. The errors for the green points are provided by the original paper [34].

5. Ground studies

Space experiments are difficult to conduct, and very difficult to control. So, several studies on ground have been performed in the frame of the ALTEA program to support space results. Ideally, the experiments should have been be similar to the ones performed in accelerators in the 1970s, with the addition of an objective descriptor of the LF, such as an EEG. However, ethical considerations did not allow anymore to put the experimenter’s eyes in the beamline in a particle accelerator. So three different experimental strategies have been conceived:

• –animal model: measuring mice EEG while irradiating one eye with short ion pulses
• –in vitro studies: studying rhodopsin activation with ion irradiation
• –human studies: studying the LF sensations from proton/hadron therapy patients reported during the irradiation.

5.1. Animal model

Mice studies permitted to demonstrate that ions in the eyes do elicit electrophysiological evoked responses. Both retinal and cortical responses have been measured and studied, showing a remarkable similarity with responses following visual stimuli.

Mice were anesthetized and one eye was irradiated with a ¹²C (200 MeV/n) beam. Very short (≈2–5 ms) pulses of ions (from 500 to 7000 ions per burst) were delivered every three seconds. Among the major results of these investigations [45], [46]:

• –electrophysiological particle evoked responses do exist, supporting the idea of using EEG as an objective LF indicator (see Fig. 5, insets, for retinal responses).
• –there is an inverted U response (large response are elicited only when the burst contains about 1600 to 2600 particles in each burst (see Fig. 4): some mechanism might inhibit the response when there are too many particles.
• –in a combined study with visual and particle stimulation [46], particle caused a magnification of the retinal responses with a concurrent decrease of the cortical ones. This might be due to a parallel path to cortex, one physiological through the visual system, one directly to optic nerve/visual cortex.

Figure 5.

The b-wave amplitude (see inset at the top right) of each electroretinogram response to a ≈2 ms burst of ¹²C delivered into the left eye of one mouse is plotted against the number of ions in the burst. The number of ions in each burst is changed not monotonically, to avoid possible fatigue or habituation biases. The different symbols indicate different times windows during the 50 min irradiation (see legend at the top left). The real time retinal responses are illustrated in the lower inset (the instant of the ion bursts is reported at the bottom, every ≈3 s). In the top right inset, the average of the largest retinal responses is shown (the peaks of these responses are indicated in the main plot with green crosses: from 15 to 45 min). The inverse U response is evident, implying a threshold effect but also some inhibition when the number of ions (dose) is too large. (from [4]).

5.2. In vitro studies

In the quest for developing a model that could explain a relevant portion of the LF, the possibility that rhodopsin (the protein at the start of the photo-electronic cascade in the visual process) could be activated by charged radiation has been investigated [47]. Rhodopsin has been irradiated with ¹²C ions in a particle accelerator. Results, based on spectrophotometric measurements, demonstrated that rhodopsin was activated by radiation, without been damaged (for doses up to about 10 mGy). Companion measurements in laboratory showed that this activation was mediated by free radicals produced by radiation in the vitreo.

These works suggested and verified the first new model for the LF generation (beside the Cerenkov effect). In this model the ions produce a large quantity of free radicals when travelling through the eye vitreo. Most of these either recombine or are quenched by the anti-oxidant largely present in the vitreo. A few, produced close to a rod (one type of retina receptors), reach the lipidic barrier around the rod before getting quenched and undergo lipid peroxidation, emitting a visual photon by chemiluminescence (Fig. 6). This photon is therefore ‘seen’ by any nearby receptor, producing the visual sensation of a LF [47], [48], [49]. The energy needed for the radicals generation is much lower than any energy released by ions, including protons, travelling through the vitreo [50]. This model, therefore, allow generation of LF mostly without concerns for the LET.

Figure 6.

Simple sketch of the rhodopsin model, mediated by free radicals produced by radiation in the vitreo, see text [47].

Finally, along a similar line of thought, a new project has been designed with the objective to verify if also calcium signals can be modulated by ion radiation in space [51]. If this were demonstrated, several brain activities could be affected by ions.

6. Proton/hadron-therapy, a new channel to study LF

The third approach for the ALTEA ground studies is slowly becoming a new important area for LF studies. Proton / heavy ion therapy, a new tool to treat cancer, exploits the peculiar way ions release energies when travelling in a tissue. While photons leave the largest amount of energy when they enter the tissue, ions do release the largest portion of their energy when they stop at a certain depth. This is the key feature to build a system to treat deep tumors minimizing the effect along the radiation entering path [52]. Space radiation in human exploration and ion therapy are the only research and technical areas where humans are subject to ion irradiation. Joined conferences are now periodically held with a continuous exchange of scientific information. Ion therapies are becoming an important asset for LF studies.

6.1. The ALTEA studies

The patients at the experimental Hadron-therapy (¹²C) center in GSI, Darmstadt, Germany were the first to report seeing LF during therapy. A first preliminary study during treatments of skull base tumors [53] clearly showed that these LF (white, with 10% of yellow ones) were appearing when the beam was on some specific region of the cortex, close to the back of the eye, possibly in the retina region. During the ALTEA investigation, further studies were performed on patients’ EEG showing, for example, that carbon ions depressed the brain alpha rhythm [4].

The first involvement of proton therapies in these studies [54], claimed that, at the Institut Curie – Centre de Protontherapie in Orsay, France, “Sixty percent of the patients treated for choroidal melanomas using 73 MeV protons report anomalous phosphenes”.

In a later paper [55], relative to therapies in Loma Linda proton center, a retrospective study on 19 patients showed, for the first time, the involvement of other three sensory channels beside visual: auditory, olfactory and gustatory. These illusion will probably require cortical mechanisms to be produced. Many of the patients perceived multiple illusions during the same therapy sessions (see Fig. 7).

This study also showed that if the dose is too large in the visual areas no visual sensations are elicited, providing also further evidence that there is an ’optimal’ dose window where LF can be elicited (‘It is possible that at higher doses there are inhibitory reactions that limit visual responses precluding monotonicity with dose’) similar to the one demonstrated in the mice study. Finally, a patient reported LF of “distinctly different colors (cobalt blue or white)” when the beam arrived from different directions during therapy.

6.2. Recent Proton therapies studies

In the last few years a number of LF studies during proton therapy has been published. An interesting discussion about LF (in therapy but also in space) can be found in [5], [56]. In these cases the most important key parameters were linked to tumor location and dose. For the detailed considerations related to these the reader is addressed to the original papers (see also [57], [58], [59], [60], [61], [62], [63], [64]). Most of the studies were prospective studies where the patient was aware of the study, and accepted it before the therapy, and was asked questions about the illusions immediately after each therapy session.

An interesting dataset from Mathis work [56] to compare with other results is the distribution of the perceived LF colors in proton therapy, illustrated in Fig. 8 left.

Figure 8.

Colors perceived during proton beam therapy (eye tumors, left panel, adapted from [56]), during heavy ion therapy (head and neck tumors, center panel, data from 53) and by astronauts (right panel, data from [39]).

The different color composition is striking and cannot be explained just by the interindividual variability reported in most the LF related studies. To these plots it should be added the value of 74% for blue LF, reported by Chuard, when choroidal melanoma were treated at the proton therapy centre of the Institut Curie in Orsay, France [60], a figure well compatible with the Mathis’ color distribution.

Color perception would require the involvement of retinal cones. Cones are in a smaller area (see Fig. 9) of the retina than rods (that show much higher sensitivity to light, but do not mediate color vision), and are mainly located in the macular area and in the fovea [5].

Figure 9.

The larger coverage of rods in the retina, with respect to cones (from [5]).

Most recently a Japanese study on adult patients [62] reported a similar color scheme as the one in Fig. 8-left, but with a lower percentage of patients perceiving light flashes. Dose to the retina appeared to be the key parameter for these illusions. In another study on pediatric patients [63] also smell perceptions were reported.

7. Discussion

The key parameters describing the LF phenomenon are numerous. The types of experiments (in space, with human in ground-based accelerators, with an animal model, in vitro, on patients) provide information that should be considered with caution especially trying to use them together. It is with this general warning that the reader should go through the following lines.

7.1. General comments

As a first comment, the quite low isotropic GCR flux implies that the probability of having two particles hitting the same structure in the eye (or cortex) at the same time (or within a reasonable Δt to be considered such) is neglectable. So a first realistic assumption is that space LF are produced by a single ion hitting a target structure, most probably the retina.

High Energy and high Z ions (HZE) are characteristics often used at the beginning of these studies to define the source. As stated, there is no need to have high Z and, furthermore, the term ‘high energy’ is probably misleading: lower energy ions deposit often more energy in matter and therefore could be more effective in eliciting LF. On the other hand, there has been evidence, since the very beginning, that bursts of low LET charged radiation may produce LF [20], and the generation model mediated by free radicals generated in the vitreo by the radiation [47], does not impose any minimum threshold for the LET.

Among the many aspects of LF, their motion did not find any reasonable explanation (this includes the absence of vertical motion). No investigation was aimed at this problem.

Another issue regarding the study of LF is the distinction between the direct and indirect mechanisms. Many authors use a physics based approach, distinguishing when the primary incident particle hits the target area (direct) from when some secondary particle, generated by some physics interaction, hits the target (indirect). In more biochemically oriented works, it may be distinguished whether the effect is directly starting the vision process (whichever particle is responsible for it) or if it is mediated by some biochemical mediator (indirect). In this second case an indirect mechanism can be mitigated by countermeasures minimizing the activity of the mediator, while a direct one only reducing the amount of radiation reaching the astronaut would be beneficial in reducing the LF rate.

Missing from all the cited works, or from any relevant discussions initiated by them, is the issue of the possible combined effect of radiation and weightlessness. For example, microgravity is a possible cause of changes in intraocular pressure (Spaceflight Associated Neuro-ocular Syndrome, SANS, see for example [65]), and this might be connected to the production of anomalous phosphenes perceptions.

Finally, from the clinical standpoint LF have been indicated as possible predictors for vision loss or therapeutic success after irradiation [56], indicating these studies as a novel field of investigation. Also, it has been mentioned that LF in space could be an indicator for free radicals in the eye (or, conversely, for the amount of antioxidant), a sort of rough biomarker. This could become an interesting point if the model mediated by radicals’ production could be demonstrated to be the most relevant one.

7.2. Long term tendency

Fig. 4 might suggest a linear decreasing relationship of LF rate with time. However, due to the extreme interindividual differences and to the many sources of uncertainties in the rate measurements (only counting errors have been considered in the error bars) no relation should be inferred. This said, I believe it would be acceptable to say that the rate of LF perception in the 1970s is higher than the rate of LF perception in the 1990s and later. This difference is difficult to explain just considering physics parameters such as solar cycle, orbit altitude, shielding. From the radicals mediated model [47], it can be conceivable that the diet might have had a role in it. In the last decades more attention at the astronauts’ diet on ground has been given, providing a larger amount of antioxidant. Also astronauts’ personal interest in the measurement could play a role in the rate of LF, however it is difficult to believe that this could explain the difference between one LF every two minutes (Skylab, ASTP) and one – two LF every hour (ALTEA).

Diets in space flights have always been more controlled and possibly richer in antioxidants. If so, this would suggest a potential explanation for the higher translunar coast LF rate that the Apollo astronauts perceived when compared to the trans-earth coast one [27]. In the first case astronauts were still under the effects of the ground diet, while in the second case mostly under the effects of the flight diet.4 Even the SilEye2 reports that LF rates were decreasing during a mission and with each subsequent flights during years [34] could find partial support from this hypothesis.

Finally, accepting the above explanation for the recent lower LF rates, and considering that diet could take care only of the excess of radicals in the eye, it could be conceived that the vast majority of LF perceived in the 1970s was generated indirectly, mediated by a biochemical agent. This is strongly suggesting further LF measuring sessions (even a simple counting experiment) in LEO to verify if these low rates are permanent.

7.3. Scale factors

In space the cause of a LF is a single particle. In all ground experiments (animal models, in vitro experiments, proton/heavy ion therapy) a much larger number of ions are involved (Table 11). In the mice experiments ≈10³ ions per burst collimated into the eye are used to irradiate the retina. In the rhodopsin experiment a rate of 10^8–9 ions/s is hitting the target.5 In the proton (or heavy ion) therapies, the largest amount of radiation is aimed at the tumor and the interactions leading to LF perceptions are most likely in the surrounding healthy tissues. These are hit only by the tail of the distribution of the beam, but, even so, we should consider at least 10^5–6 particles per second. In most of the therapy works on LF, 0.1–1% of the maximum dose, is considered ‘no dose’, and this, can be really misleading for LF considerations. As an example in a study on extraretinal induced visual sensation (photon therapy) it was enough to have less than 1% of the dose on the retina (daily fraction of 2 Gy) to consider the retina not irradiated when studying LF perception [66].

Table 11.

The scale problem, number of ions involved in different experiment.

Type of measurement	Number of ions involved
Space	1
Ground works in the 70s	1–103
Mice	103/s*
Proton / heavy ion therapy	105–6/s
In vitro	108–9/s

* delivered in burst of about 2 ms.

This scale issue might also have effects on the way LF are perceived. The perception of LF in orbit is obviously in sync with the particle passages, as well as for the electrophysiological responses in mice (that are synced with the particle burst arrival). This feature has not yet been reported during therapies, with one exception [55]. That is probably due to the way the beam is delivered, and to the absence of a means to measure the possible synchronicity, with the large number of ions possibly hitting concurrently the sensible areas.

7.4. Efficiency

Many researchers attempted to provide an evaluation of the radiation efficiency in producing LF as indicated in Table 12.

Table 12.

Estimated efficiency (LF per ion) as evaluated in the different experiments.

Study	Estimated efficiency (LF / particle in the eye)	Note	Reference
Space – SilEye2	1.3 × 10−2	Z > 1	[34]
	1.7 × 10−5	Z = 1
Space – ALTEA	2 × 10−3	H < 45 MeV	[4]
Ground – Mice	1 × 10−3	12C 200 MeV/n – collimated	[45]

In the SilEye2 estimation the protons were considered separately, as the authors claim that two different efficiencies (for protons and for Z > 1 nuclei) should be calculated. However, in that paper the anisotropy of the SAA flux (directional) was not considered, and this is why the stated efficiency for protons should be taken with caution. In the ALTEA measurements most of the protons were excluded. If we take into account all the protons, efficiency might go down by 1–2 orders of magnitude, well below the level of the Mice experiment.6

The efficiency evaluations involve several parameters. The simplest is linked to a geometrical cross section. As an example, the ≈10³ ions to elicit a response in a mouse may be interpreted as if only one out of one thousand hit the right target in the eye. Another parameter could involve the probability of the physics interactions giving rise to that specific particle that could produce the sensation. A third could describe a similar probability linked to the possible biochemical behavior.

Interestingly, the estimated efficiencies appear all within one order of magnitude, and possibly suggest that a simple geometrical cross section might play the key role.

Finally, during therapies the efficiency parameter does not have a real meaning: either the patient perceives LF or not, and this is likely linked to the scale problem (7.3) and to the large amount of ions the sensible target receives in the unit of time.

7.5. Models

For all fast ions, Cerenkov radiation is a possible source of LF. It may be seen as a quasi-physiological mechanism as the photon, directly produced by the radiation, is generated inside the eye instead of outside, but from there on it follows the physiological path. Recently Cerenkov light has been measured from the eye of patients undergoing radiotherapy [67].

All proton and heavy ion therapies are performed with particles with energies lower than the threshold for Cerenkov effect. However, it has been recently suggested that Cerenkov radiation could be indirectly elicited by protons at energy lower than the threshold by emissions of fast electrons liberated by prompt gamma and neutron emission or by the decay of radioactive positron emitters, even if with very low probability [68].

Light flashes can also be produced indirectly by a biochemical mediated process [47], [48], [49] or, more in general, through biophotons generated by overproduction of free radicals from ionizing radiation in the retina. [69], [70], [71].

The radicals mediated process does not have a low LET threshold, and this could be a welcomed characteristic to describe proton induced LF.

Finally, the radicals mediated model might lead to an inverse dose relationship. The radicals in the eye have to reach their target (lipid barrier around the rod) before they recombine with each other or get quenched by antioxidant. If the flux of ions increases, the recombination process is more likely and the dose relationship may invert. This would explain the inverse U relation measured in mice [45] and the peculiar inverse dose relationship at high doses, demonstrated in the Loma Linda proton therapy patients [55].

Direct retinal excitation involving intrinsically photosensitive retinal ganglion cells have been also proposed as an LF starter [5].

7.6. Cortical interactions

The existence of illusions in senses other than visual [26], [55], [61], [63], [64], the inverse effect on electroretinogram and on cortical evoked potentials following concurrent light and ion stimulation [46] strongly suggest cortical interactions and therefore requires models that are not relying just on the retina. These models are still lacking, and further studies should be promoted in this sense.

7.7. Colors

The color perception (Fig. 8) in space has a large white predominance, and it is like the color perception during hadron therapy when head and neck tumors were treated. The different distribution of color perception in proton therapy when treating eye tumors, with a colored large predominance, supports the idea that, to produce color LF, the fovea and the cones need to be hit. Indeed, the fovea is small (Fig. 9) and rarely hit in space (isotropic radiation) and while treating head and neck tumors (radiation mostly directed to a different area). Conversely, during eye tumor proton therapy, the beam aims directly at the eye and the probability to hit the fovea is higher. This explanation could also support the report of different color (white and blue cobalt) in proton-therapy when the beam entrance was different, it could well be that in one case the beam hit the fovea and the LF was therefore colored (blue cobalt). Finally, even during the Moon flights, again with isotropic radiation, the perceived LF where white (with a single exception), supporting this suggestion.

7.8. South Atlantic Anomaly

The LF rate while passing over the SAA, in the LEO measurements has always been a reason for debate. The very first measurement, in the Skylab 4 mission, 1974, reported a very high rate of LF. Both sessions did report this sharp increase, even if the first one more moderate then the second one. The immediate comparison with the following experiment one year later during the ASTP program where no increase was perceived has been the core of many debates. Physical parameters could explain it partially (different altitude and shielding), and some other explanation was needed to account for the full difference. As noted, the directionality of the SAA radiation could be a key point in this, the highest LF rates being perceived when a specific effort was put to align the Earth magnetic field lines with the anterior – posterior axis of the astronaut’s head.

Beside interindividual variability, other explanations for this discrepancy were also postulated, such as nuclear interactions close to the retina like nuclear star formation [72].

However, a striking issue is that since the very first report (Skylab 4), no one else did report a significant increase of LF rate during SAA passages (ASTP, SilEye, SilEye2, Alteino, ALTEA). This might suggest a lower capability of protons to induce LF, but the evidence brought by the results coming from proton therapy does not support this. Again, directionality of the radiation may play a role, as well as the apparently lower sensitivity for LF perception during the last decades (Fig. 4).

7.9. Future

The many investigations on LF provided answers to many questions, but as shown, some remain unanswered and several new questions have been posed and might require further investigations.

Experiments to confirm the low LF rates measured in the last decades in Fig. 4 would be welcomed. Even simple counting experiment would provide more ground for claiming that the astronauts diet might have improved their resilience to radiation, through mitigating the radicals effect. This might include also repeating a questionnaire-based study to understand if the LF perception incidence decreased with respect to the results published previously [39]. If possible, the LF rate correlation with the diet should also be directly investigated.

Motion of LF is still not explained. Studies directly aimed at this issue can clarify the overall panorama of the involved mechanisms.

Prospective studies on patients (proton therapy or heavy ion therapy) could improve our knowledge about the extension of these effects over all the sensorial pathways. On the same line, works on interaction mechanisms that could be effective also for illusion other than visual should be carried out.

8. Conclusions

The interest on the LF phenomena started with an operative question related to the risk for those astronauts perceiving such illusions. Now this issue seems to have less importance, especially due to the many humans who flew in space not showing any pathology linked to LF. However, it really is still a partly unsolved puzzle, becoming an interesting research issue which may still have significant weight in the development of radiation countermeasures for space travel. While about source (cosmic radiation) and some of the targets (eye, retinal structure at least) there is consensus, the interaction(s) starting the perception process are still under debate and new data is needed. It is also easily conceivable that LF investigations could provide information about if and how single particles in space could interact with brain function, even if sporadically and rarely, impacting with the crew ability to optimally perform a mission.

Declaration of Competing Interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.