Diving.

Following the onset of each dive, [tHb] declined rapidly. The date of decline in [tHb] slowed during the descent phase of a dive (Fig 2B). Rate of decline in [tHb] remained relatively constant throughout the remainder of the dive, reaching a minimum approximately 10 s before surfacing. [Hb_diff] showed an initial increase for ≤10 s (±2 s) during the descent phase, reaching a maximum and TSI oxygenation maximum for each dive. [Hb_diff] then declined monotonically throughout the remainder of the dive (Fig 2). Towards the end of a dive, about 10 s before surfacing, [tHb] increased rapidly, whereas the rate of [Hb_diff] decline increased and [Hb_diff] reached a minimum. [Hb_diff] during diving remained below prediving levels.

Surface intervals.

On surfacing, [tHb] continued to increase rapidly, reaching a maximum in the second half of the surface interval (Fig 2B). [Hb_diff] and TSI increased steadily throughout this period. Maximum [tHb] during surface intervals was greater than prediving [tHb] levels, peaking at 62% ± 15% of the way through the surface interval. During the second half of a surface interval, [tHb] declined rapidly, whereas [Hb_diff] and TSI continued to increase. Of the 102 dives, 30 did not display a reduction in [tHb] until after the start of the next dive. Of these 30 dives, 70% came from one animal—Thor (S1A Fig).

Postdive resting period.

Following experiments, postdiving [tHb] remained higher than prediving levels throughout the resting period for at least 5–15 min. [Hb_diff] did not reach prediving levels until 5–15 min of the postexperiment rest period.

Diving.

Following the onset of each dive, [tHb] increased, reaching a maximum 20 s (± 4 s) into the descent phase (Fig 4B). Throughout the period of increasing [tHb], [Hb_diff] showed an initial increase during approximately the first 5 s of the descent phase, followed by a short rapid decline ending 20 s (± 4 s) into the descent phase (Fig 4B). Over this time course, TSI dropped rapidly. Following [tHb] maxima, [tHb] fell for the remainder of the descent phase and continued falling when the animals were inactive at the feeder (Fig 4B). This [tHb] decline was characterised by a greater reduction in [HHb] than in [O₂Hb] (Fig 6B), and in one seal in fact [O₂Hb] increased. As [tHb] declined, [Hb_diff] showed a synchronous increase, forming a secondary peak (Fig 4B, dashed red line) approximately 50 s into each dive. Consequently, TSI increased rapidly throughout the period of secondary reoxygenation and falling [tHb]. On some dives, the secondary [Hb_diff] peak represented a dive cycle [Hb_diff] maximum—i.e., [Hb_diff] during a dive was greater than at the beginning of the dive (Fig 4). After [tHb] minima and commensurate secondary [Hb_diff] peak, [tHb] increased steadily while [Hb_diff] and TSI declined steadily throughout the remainder of the dive until the animal surfaced. Cerebral [Hb_diff] during diving remained below prediving levels.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Cerebral haemodynamics (‘Haem’) of O₂Hb and HHb for two seals during two key phases of a dive.

(A) Mid-surface interval and (B) early in a dive when secondary reoxygenation occurred. Dots represent median concentration change. Error bars represent the interquartile range. Changes are shown as relative changes from CBV minima during a PDSI. Dives by Vebjörn, n = 23. Dives by Ulf, n = 54. Data can be found here: https://doi.org/10.5061/dryad.k67cg66. CBV, cerebral blood volume; HHb, deoxyhaemoglobin; O₂Hb, oxyhaemoglobin; PDSI, postdive surface interval.

https://doi.org/10.1371/journal.pbio.3000306.g006

In the diving trial shown in Figs 4 and 5, the last dive shows considerable, rapid reduction in [tHb] (Fig 4A), likely in response to human presence outside of the breathing chamber, which the seal did not expect. During the preceding surface interval, the seal appeared reluctant to surface, spending <5 s with the nostrils above the surface before crash diving and swimming rapidly away from the breathing chamber.

Surface intervals.

On surfacing from a dive, cerebral [tHb] declined until after the midpoint of the surface interval (58% ± 0.2% of the surface interval) (Fig 4; S1 Fig). Conversely, [Hb_diff] and TSI increased steadily throughout the first half of each interdive surface interval (Fig 4A). During some dive cycles, this fall in [tHb] during the postdive surface interval (PDSI) resulted in minimum [tHb] within a dive cycle. This pattern of reduction in cerebral [tHb] occurred in 71 of the 77 dives for cerebral NIRS measurements. Throughout the remainder of the surface interval, both [tHb] and [Hb_diff] increased (Fig 4B, solid red box), whereas TSI decreased. Increased [tHb] was characterised by an initial increase in median [O₂Hb], followed by a subsequent increase in [HHb] (Fig 6A).

Postdive resting period.

Following experiments, postdiving [tHb] remained higher than prediving levels throughout the resting period for at least 5–15 min. [Hb_diff] did not reach prediving levels until 5–15 min of the postexperiment rest period.

Changes in advance of diving.

An increase in cerebral blood volume (CBV) occurred in advance of diving, representing a ≤3.2% increase in CBV (see S1 Text). The increase in CBV that occurred throughout the second half of each surface interval (beginning 58% ± 0.20% of the way through the surface interval) and the first 20 s of each dive (Fig 4) resulted from an increase in [O₂Hb] and [HHb] over the first 15 s of this period of increasing CBV (Fig 6A) and thus in [tHb] (Fig 4). The net result of these increases was decreased TSI (Fig 4). Decreasing TSI, concomitantly with increasing [tHb], indicates a greater relative increase in [HHb] than [HbO₂]. A greater relative increase in [HHb] is consistent with a mismatch between cerebral inflow and outflow, whereby the rate of inflow of blood into the brain exceeds the rate of outflow, and/or greater oxygen removal by cerebral tissue than can be met by delivery.

An increase in arterial pressure would explain the initial increase in cerebral [HbO₂] in Fig 5B, as increased arterial pressure will increase cerebral blood flow and thus the delivery of oxygenated arterial blood to the brain [37]. If increased arterial delivery was matched by increased venous drainage, cerebral oxygenation would increase. However, if increased delivery of arterial blood is not matched with increased venous outflow, then cerebral oxygenation will decrease as the venous contribution of the intracerebral blood volume increases [37,38]. Increased central venous pressure inhibits venous outflow, thereby creating a mismatch in inflow and outflow from the brain, leading to pooling in the venocapillary system and decreased cerebral oxygenation [39–41]. We, therefore, hypothesise that decreasing BBV, which begins in advance of diving (and the likely onset of bradycardia), is indicative of increased peripheral restriction, which leads to increased arterial pressure (increasing cerebral blood flow to the brain). The resulting increased central venous pressure inhibits venous outflow, causing venous congestion, which can explain the greater relative increase in [HHb] as seen in Fig 6A. The consequences of increased arterial and central venous pressure would act synergistically to increase CBV and reduce cerebral oxygenation [37–43] and would explain why [tHb] increases yet TSI decreases during the second half of each interdive surface interval.

Seemingly concurrent with the increase of CBV during the second half of an interdive surface interval, BBV rapidly declined. A decrease in BBV in advance of diving is consistent with the onset of arterial restriction associated with the dive response [4], whereby blood flow is reduced to peripheral tissues nonessential to diving in order to conserve oxygen. As the onset of the reduction in BBV, in advance of diving, occurs 16.14 ± 8.00 s before diving, the onset of peripheral restriction appears to commence well in advance of diving bradycardia, which commences at the point of submersion [1,9]. Even if there was anticipatory bradycardia, which has been demonstrated in harbour seals [18], the onset of change in BBV in advance of diving in the current study almost always exceeded published durations of such anticipatory bradycardia (<2 s before diving). It would, therefore, follow from this reasoning that there is a period during the second half of each surface interval during which there is a mismatch between CO and arterial resistance. When CO exceeds the level of arterial resistance, arterial and venous pressures will elevate [13]. Indeed, in experiments in which CO has been artificially uncoupled from increased arterial sympathetic tone by inhibition of diving bradycardia (either by cardiac pacing [17], efferent vagal blockade [44], or atropine treatment [45]), arterial and venous pressures show a marked rise. At the level of individual tissues, such as the brain, increased arterial and venous pressures have the effect of both increasing blood volume and decreasing oxygenation [39–42], both of which were observed in the current study.

Decreasing cerebral oxygenation in advance of diving seems undesirable. However, if (as we hypothesise) the decreasing TSI is driven by intravascular pressure changes associated with the onset restriction, and cerebral deoxygenation can be tolerated, which it appears seals can [21–23], then decreasing cerebral TSI would allow seals to begin a dive in an already reduced oxygen consumption state. If this were correct, decreased TSI in advance of diving would not in itself be an anticipatory response but rather a physiological by-product of systemic perfusion changes.

An alternative explanation for the increase in CBV and reduction in TSI during the second half of each surface interval could be an increase in cerebral metabolic rate in which there is greater oxygen removal by cerebral tissue than can be met by delivery. This could explain both increased CBV and the relatively greater increase in [HHb] than [HbO₂]. Increased cerebral metabolic rate, certainly during cortical activation, has an associated haemodynamic epiphenomenon in which there is arterial vasodilation and increased delivery of oxygenated blood. The result is that active cortical regions are characterised by increased [HbO₂] and [tHb] but reduced [HHb] because the rate of blood oxygen delivery exceeds demand. This hemodynamic signal is the principal underpinning both blood oxygen level–dependent (BOLD) MRI and functional NIRS (fNIRS) bioimaging techniques [28]. We suggest, however, elevated cerebral metabolic rate is a less likely explanation for increased CBV in advance of diving than changes in intravascular pressure. First, the magnitudes of change in [tHb] during the hemodynamic epiphenomenon associated with increased cerebral metabolism are ≤1 μ.mol.L⁻¹, and the changes in the present study are much greater (~5 μ.mol.L⁻¹). Second, the epiphenomenon associated with elevated cerebral metabolic rate supports oxidative metabolism by ensuring that delivery of blood oxygen exceeds metabolic demand; hence, TSI increases. Yet, in the current study, TSI decreases, which is inconsistent with cerebral vasodilation associated with elevated cerebral metabolic rate. Admittedly, this pattern of reduced TSI could be a consequence of unique cerebral reconfiguration in seals [23]. The hypothesis of elevated cerebral metabolic rate in advance of diving and its consequence on blood volume and oxygenation could be tested using NIRS with inclusion of additional wavelengths to measure concentration of cytochrome c oxidase, which would capture changes in cerebral metabolic rate specifically and not just [HbO₂] and [HHb] as in the present study.

Secondary reoxygenation.

Seals abruptly reoxygenated cerebral tissue approximately 20 s into each dive, indicated by a marked increase in TSI (Fig 4A). During this ‘secondary reoxygenation’ phenomenon, declining [tHb] (Fig 4B) was dominated by a reduction in [HHb] (Fig 6B). Changes in intravascular pressure are again suggested here as the most probable explanation for the period of secondary reoxygenation. Following the onset of diving and bradycardia, it would be expected that adjustments in peripheral resistance, HR, and CO should allow arterial and central venous pressures to return to normal after the proposed increase during the latter part of the PSDI described above. Decreased central venous pressure should then promote increased venous outflow from the brain and thus reduce cerebral venous pooling [39–41]. Reduced venous pooling would simultaneously decrease CBV and increase TSI as a result of increased [HHb] removal. Decreasing the venous contribution of CBV would then increase cerebral TSI without any requirement for changes in systemic oxygenation. This explanation is consistent with the patterns here of decreased [tHb] early in each dive and increased TSI despite a lack of access to ambient air. Arterial–venous contributions within a tissue’s vascular bed are an important aspect of tissue-specific oxygenation in that tissue oxygenation relies not only on systemic blood oxygen levels but also on tissue-specific perfusion [40]. It is important to assess cerebral oxygenation directly rather than to infer oxygenation from systemic blood gas values [40].

While we do not have any blood pressure data to support our hypothesis for changes in arterial and/or venous pressures, we propose intravascular pressure changes as the explanation underpinning and ultimately linking the opposing patterns of change in CBV and cerebral TSI in advance of and shortly after each dive. Intravascular pressure measurements [46] and HR collected simultaneously with NIRS data would help support or refute our hypotheses.

An alternative explanation for the secondary reoxygenation phenomenon could be changes in systemic blood oxygen levels. The capacity for seals to infuse a bolus of oxygenated blood from the spleen into circulation during diving has been shown in freely diving Weddell seals and northern elephant seals (Mirounga angustirostris) [20,47]. A bolus of oxygenated blood resulted in an increase in haemoglobin (Hb) content in the major arterial vessels [20]. Contradictory responses of [tHb] and [Hb_diff] in the current study suggests that a similar infusion of oxygenated blood from a central venous reservoir cannot fully explain the repeated pattern of cerebral oxygenation seen here. In the present study, there was a decrease in [tHb] in both the brain and blubber during secondary reoxygenation rather than increased [tHb], which could be expected if systemic Hb concentrations were elevated following infusion of previously sequestered blood. A lack of a secondary reoxygenation pattern in blubber also indicates the cerebral reoxygenation is not the product of systemic [HbO₂] changes. Furthermore, injection of splenic stored blood cells is not a plausible explanation for the repeated reoxygenation across long sequences of dives. The short interdive surface intervals here (43 ± 8 s) make it unlikely that there would be sufficient time for seals to reoxygenate and replenish the splenic reservoir in time for repeated infusion across multiple dives. If a central venous reservoir is involved, it seems likely to comprise the hepatic sinus and caval vein [32].

A second alternative explanation for reduced CBV during the secondary reoxygenation phenomenon early in each dive could be a result of vasoconstriction of selected vascular beds in the brain. Blix and colleagues [4] showed reductions in brain blood perfusion in the early phase of the dive. Vasoconstriction would effectively reduce the cerebral vascular bed and thereby also reduce CBV. However, vasoconstriction would neither elevate TSI or [Hb_diff] nor explain the greater relative reduction in [HHb] in the current study. We, therefore, suggest that changes in intravascular pressures are a more probable explanation for secondary reoxygenation accompanied by a simultaneous decrease in CBV.

Remainder of the dive.

After the secondary peak in [Hb_diff] and coincident [tHb] minimum for a dive, associated with secondary reoxygenation, CBV increased gradually throughout the remainder of a dive, whereas [Hb_diff] showed a monotonic decline (Fig 4). Increasing CBV later in the dive may reflect localised hyperaemia in response to decreasing blood O₂ and increasing CO₂ [38] and demonstrates a similar perfusion pattern as shown by Blix and colleagues [4], as determined using microspheres. CBV declined rapidly on surfacing and throughout the first half of the surface period, consistent with rapid cerebral vasoconstriction, coincident with rapid cerebral reoxygenation (Fig 4) as seals reloaded depleted blood oxygen stores and off-loaded CO₂ during postdive recovery. Alternatively, the decline of CBV on surfacing could be associated with the reduction in peripheral restriction [4].

As other authors have demonstrated, we show that seals do not fully reoxygenate Hb to prediving levels between repeated dives with short surface intervals, resulting in a downward trend in [Hb_diff] over consecutive dives (Fig 4) [48]. Coincident with diminishing [Hb_diff] at the start of successive dives, the magnitude of secondary oxygenation increased with each dive in a sequence (Fig 4). Increasing secondary reoxygenation in response to declining starting oxygenation could facilitate short surface intervals while maintaining adequate oxygen availability to the brain while submerged. The resulting shortened interdive surface intervals may allow seals to maximise time spent foraging and more fully exploit a transient prey patch while maintaining adequate oxygen delivery to key tissues that would otherwise be possible only if the seal remained at the surface until Hb was fully saturated.

Following the end of a diving bout, there was a protracted period of elevated [tHb]. Elevated [tHb] may be indicative of postexercise hyperaemia associated with recovery. Blix and colleagues [4] found similar increases in blood following diving and proposed it was a response to postdive recovery.

Limitations of method.

Due to the novel application of NIRS in the current study, a number of assumptions had to be made about the optical properties of seal Hb and tissue. First, we have relied upon the fact that Hb of northern elephant seals has absorption peaks that are identical to cattle, sheep, and multiple other mammalian species, including humans [49], to assume that the properties of harbour seal Hb are also the same as human and elephant seal Hb. We argue that this is a reasonable assumption given the identical nature of Hb from a variety of mammals. Second, the absorption coefficient and reduced scatter coefficient for seal tissues are not yet published, and we therefore had to assume a differential pathlength factor (DPF) of six—equal to that used for human cerebral measurements as well as adult sheep [50]. Solving for the optical properties for seal tissues may affect the exact relative changes in micromolar units per litre of tissue for [HbO₂] and [HHb] and therefore [tHb] and [Hb_diff], too. A change in DPF of 1.0 will change the magnitude of change in [HbO₂] and [HHb] by 20%, and a change in DPF of 2.0 changes the magnitude by 40% [51]. However, as the key focus of this manuscript is the patterns of change in tissue blood volume and oxygenation and the timing of these changes across repeated dive cycles, the conclusions and key results would be unaffected by a change in DPF.

NIRS system.

An animal-borne archival NIRS sensor, PortaSeal, was developed from an existing wearable dual-wavelength CW-NIRS system for humans (PortaLite mini) that simultaneously uses the modified Beer-Lambert law [53] and SRS methods. Changes in the concentration of HbO₂ and HHb can be calculated from changes in light absorption using a modification of the Beer-Lambert law, which describes optical absorption in a highly scattering medium [54]. Since the absolute concentration of chromophore is unknown, relative values for HbO₂ and HHb were measured as a change from the start of the first dive following a 5–15 min prediving period in micromolar units per litre of tissue (μ.mol.L⁻¹). This allowed calculation of total Hb ([ΔtHb] = [ΔO₂Hb] + [ΔHHb]) that can be used as a proxy for changes in blood volume and calculation of relative Hb difference ([ΔHb_diff] = [ΔO₂Hb] − [ΔHHb]). Measures of [HbO₂] and [HHb] allowed calculation of [tHb] and [Hb_diff]. [tHb] provides information on Hb volume without the influence of relative oxygenation changes ([Hb_diff]). [Hb_diff] provides information on relative Hb oxygenation without the influence of changes in Hb volume. TSI is a measure of oxygenation (measured using a theoretical photon propagation model in a highly scattering medium [51,55]) that provides a measure of the collective effect [tHb] and [Hb_diff] on tissue oxygenation. NIRS measures a volume of tissue containing microvasculature, comprising capillary, arteriolar, and venular beds but not arteries and veins. NIRS is weakly sensitive to blood vessels >1 mm in diameter because they completely absorb the light [28]. CW-NIRS has previously been used to measure [HbO₂], [HHb], and [tHb] in nonhuman species such as sheep [50] and pigs [56] and TSI in species such as dogs [57]. Seal Hb has the same optical properties as human Hb [49], allowing the previous application of invasive, muscle-implanted NIRS in seals [25,26]. However, the absorption coefficient and reduced scattering coefficient for seal tissues are not yet published. Therefore, a DPF of six has been assumed for all measurements [58]. The PortaLite mini consisted of a sensor body, housed in an aluminium case with a removable O-ring sealed lid, and a sensor head. To waterproof the sensor head, the light diodes, photodiode receiver, and printed circuit boards (PCBs) of the sensor head were first fitted into optically opaque polyoxymethylene housings. The housings were filled with spectrally transparent epoxy (EPO-TEK 301, Epoxy Technology, Billerica, MA, United States). To ensure that epoxy encapsulated the electronics but did not cover the optical window on the optodes, the sensor head was cradled in a custom-built silicone mould. This allowed the internal components to be filled and waterproofed by epoxy but ensured the external surface and LEDs remained exposed. Once the housings were filled with epoxy, an optically opaque lid was placed onto the housing and fixed with two screws. Finally, the exterior of the housings and interoptode were potted in optically transparent epoxy for additional robustness and waterproofing. This solution for waterproofing ensured that light intensity was not impacted by the waterproofing process. Distance between light sources and detector were 28 mm (850 nm and 751 nm optode), 33 mm (852 nm and 751 nm optode), and 38 mm (851 nm and 752 nm optode), allowing optical penetration to three different depths (larger spacing between the optode and receiver allows deeper optical penetration; related to Fig 1). The response time of the PortaSeal was ≤0.1 s.

Seal diving trials.

Diving trials were conducted with four temporarily captive harbour seals in a purpose-built foraging pool at the Sea Mammal Research Unit, University of St Andrews [52]. The PortaSeal was attached to a seal prior to each recording session during brief manual restraint of <1 min. The sensor body was attached to a baseplate superglued (Loctite 4861, Henkel, Winsford, United Kingdom) to the fur at the base of the skull. The sensor head was inserted into a photopolymer chassis superglued (Loctite 4861, Henkel, Winsford, UK) to the fur on either an animal’s head or shoulders and covered with an aluminium plate to maintain contact between the optodes and skin. An area of fur equal to the inner area of the chassis was trimmed to the level of the skin using rabbit-nosed clippers. The seal was then held for 5–15 min in a holding pen with access to ambient air to gain prediving NIRS measures.

A diving session started by opening an underwater gate in the pen, allowing access to the simulated foraging setup in a covered 32 × 6 × 2.4 m pool. The seal was then free to make repeated swims between a breathing chamber and a feeding station [52] (S3 Fig). An arrangement of lanes allowed the feeding station to be set at 58 m swimming distance from the breathing chamber. The seal could only surface in the breathing chamber but was allowed to dive and surface ad libitum. An experiment ended once the seal had taken a predetermined (specific for an individual’s body mass) amount of fish at the feeding station. The seal then remained in the holding pen for 10–30 min to acquire postdiving spectrographic measurements. The PortaSeal was then removed and data downloaded to a PC.

NIRS system location.

For cerebral measurements, the sensor head was located over the parietal lobe of the right hemisphere, approximately 2 cm from the longitudinal cerebral fissure. Location of the parietal lobe was determined by locating three anatomical features—the sagittal crest and the posterior ridge of the parietal and the caudal ridge of the eye socket. For blubber measurements, the sensor head was located on the ventral surface of the shoulder region. Skin, blubber, and skull depths at each sensor head location were measured, prior to photopolymer chassis attachment, with an ultrasound scanner (CS-3000, Diagnostic Sonar, Livingstone, UK).

Tissue depth.

Depth from the skin surface to the brain was 8 mm in all animals (including 2 mm skin depth and 1 mm skull depth). Depth of blubber on shoulder region ranged from 17 to 20 mm (including 2 mm skin depth).