Electrophysiological Signature Reveals Laminar Structure of the Porcine Hippocampus

Morphology of porcine hippocampus. Sagittal and coronal sections of porcine brain were stained histologically to show organizational structure. A, Coronal (top) and sagittal (bottom) sections stained with LFB/CV show white and gray matter composition in dorsal hippocampus, which can be accessed stereologically from the brain surface. Both coronal and sagittal cuts through hippocampus produce arrow-like structures with consistent anatomic layers (scale = 2 mm). B, Coronal section stained with CV showing gyrencephalic structure (bregma –1.5mm). Deep brain structures including hippocampus (dorsal and ventral) and thalamus are visible (scale bar = 2 mm). C–E, Hippocampal layers including the alveus (A), stratum oriens (O), stratum pyramidale (P), stratum radiatum (R), stratum lacunosum-moleculare (L-M), stratum moleculare (M), and stratum granulosum (G) are clearly visible (scale bar = 100 μm). C, Staining with LFB shows the myelinated axons, while counterstaining with CV highlights neurons in the pyramidal and dentate cell layers. Columns on the left show the width of corresponding hippocampal layers quantified from H&E-stained sections in Yucatan pigs (bregma –1.5 mm; n = 8) and Long-Evans rats (bregma –4.2 mm; n = 6). The widths for the pig hippocampal layers are (in μm, mean ± SEM) A = 173.8 ± 11.55, O = 178.3 ± 15.01, P = 213.2 ± 15.9, R = 394.1 ± 22.36, L-M = 223.1 ± 11.7, M = 335.1 ± 17.85, G = 81.3 ± 4.5. The corresponding widths for rat hippocampal layers are (in μm): A = 69.1 ± 9.05, O = 144.4 ± 7.51, P = 53.1 ± 1.9, R = 343.9 ± 12.33, L-M = 112.2 ± 8.21, M = 232.8 ± 11.35, G = 90.6 ± 3.52 (scale bar = 200 μm). The following hippocampal layers were significantly larger in pigs than in rats: stratum alveus by 152% (p < 0.001), pyramidale by 301% (p < 0.001), radiatum by 15% (p < 0.05), lacunosum-moleculare by 99% (p < 0.001), and stratum moleculare by 44% (p < 0.01), whereas the stratum oriens and granulosum were not significantly different. D, Staining with MAP2 identifies location of dendritic arbors, which are densely packed. E, A subpopulation of interneurons is identified using parvalbumin (PV) IHC in pyramidal and dentate granule cell layers. F–H, The pyramidal CA1 layer (scale bar = 50 μm). F, The pyramidal CA1 cell layer is widely dispersed, similar to the human CA1 layer, as shown with LFB/CV staining. G, Pyramidal CA1 cells send their dendrites down toward the hippocampal fissure as highlighted with MAP2 staining. H, PV⁺ interneurons located above, inside, and below pyramidal CA1 layer send their projections into stratum radiatum.

Single-unit recordings in porcine hippocampus corroborate laminar structure. A, A representative coronal section from the MRI atlas (bregma –1.5 mm) within the 12-mm AP range of interest shows the orientation of deep structures such as dorsal and ventral hippocampi (Saikali et al., 2010). We selected ML 7 as an access plane to the widest part of dorsal hippocampus. B, A sagittal section through ML 7 taken from the MRI atlas shows position and approximate size of the dorsal hippocampus. The image was rotated so that cortex above hippocampus is in the horizontal plane (horizontal dashed line). An approximate size of the hippocampus is shown between vertical dashed lines from right (AP 0) to left (AP –10). C, Photograph of a pig skull in situ in the novel stereotaxic instrument with an arrow pointing to bregma and a circle indicating a craniectomy site. The main parts of the stereotaxis are frame base (1), bite bar (2), anterior-posterior (AP) bars (3), and pins (4). The animal’s head is first guided between AP bars and onto an adjustable bite bar, then a set of four pins is inserted into the zygomatic bones. Location and size of the craniectomy are shown (black circle). D, ML coordinates obtained with this stereotaxis are fairly accurate, but AP coordinates are not. Therefore, electrophysiological mapping is needed for each experiment. The digitized tungsten map from our initial animal is shown in the AP plane at ML 7 overlying a representative sagittal section, stained with LFB/CV (scale bar = 2 mm). The lines outline deep brain structures including subcortical white matter, ventricle, hippocampus, and thalamus. E, A representative sagittal section from the initial animal stained with H&E (scale bar = 1 mm) contains Hamilton syringe tracks in the AP plane in 2-mm steps from AP 0 to AP –10 (dashed lines). F, The precision and reproducibility of our targeting technique in 13 animals was estimated histologically in the ML and AP planes and electrophysiologically in the DV plane. The estimated distances are (in mm): ML = 5.91 ± 0.17 from midline, AP = –1.60 ± 0.16 from bregma, and DV = 18.76 ± 0.57 (mean ± SEM, n = 13) from brain surface to first multiunit activity. G–I, Example of single spiking activity during the mapping procedure performed by slowly advancing a tungsten electrode along the indicated track. G, A coronal section of the dorsal hippocampus stained with LFB/CV is shown from a subsequent animal from which tungsten electrode recordings were made during the initial mapping procedure (scale bar = 100 μm). Arrow points to the subsequently inserted silicon probe artifact. Histologically identified layers are labeled: alveus (A), stratum oriens (O), stratum pyramidale (P), stratum radiatum (R), stratum lacunosum-moleculare (L-M), stratum moleculare (M), stratum granulosum (G). Three-minute recordings were made at regular intervals spaced 100–200 μm apart. Raw neural signals were processed offline, and putative single units were isolated. Depths at which single units were successfully isolated are indicated by black and white circles. Representative single units from each hippocampal layer are shown as averaged waveforms (mean ± SD) accompanied by firing rates, autocorrelograms, and interspike interval histograms (Robertson et al., 2014). H, To capture all multiunit activity and confirm the locations of the cellular layers, the RMS power of the high-frequency bandpass filtered signal (600–6000 Hz) was calculated. A representative profile of spiking activity shows several peaks at the corresponding pyramidal and granular cell layers. I, Firing rates of single units recorded in 2 animals with tungsten electrodes were averaged within each of 8 layers spanning ∼200 μm each, corresponding to the hippocampal layers calculated in Fig. 1C (mean ± SEM, n = 2).

Single units recorded with silicon laminar probes confirm laminar structure. A, Photograph of the custom-made Edge silicon probe shows an internal reference and linear orientation of its 31 channels. In each animal, this probe was inserted into the dorsal hippocampus based on initial electrophysiological mapping. Electrical activity (B–D) is shown as recorded from the final position of the probe. B, Example in one animal of the laminar profile of single-unit activity shown as averaged waveforms (mean ± SD) aligned to the 31 recording channels of the silicon probe along the vertical axis (200 μm spacing), indicated by the gray cartoon of the probe on the left. Lowercase labels (a–e) indicate single-unit examples displayed in C. C, A coronal section of the dorsal hippocampus stained with LFB/CV is shown from the same animal (scale bar = 100 μm). Arrow points to silicon probe artifact, at approximately midshaft. Open circles indicate channels on which single spiking units were isolated revealing a large cluster in the region of stratum pyramidale and a single channel in the region of stratum granulosum (scale bar: x-axis, 2 ms; y-axis, 100 μV). Examples of isolated spike displayed as averaged waveforms (mean ± SD) are shown along with autocorrelograms and interspike intervals histograms (Robertson et al., 2014). Alignment of single-unit examples to the histology was based on recovered artifact of the probe tip (not shown), stereotactic coordinates used during insertion, and known spacing of the 31 contacts. Vertical alignment is only approximate, as the plane of insertion was oblique to the histologic coronal plane and the relative thicknesses of the hippocampal layers change along the AP axis. D, Stratum pyramidale of CA1 was identified in four animals recorded with the Edge laminar silicon probe based on the maximum power of multiunit spiking activity (RMS power in the 600–6000-Hz band; mean ± SEM, n = 4). The traces from all four animals aligned at the peak of the RMS power are shown. Only aligned channels where all four animals overlapped are included (total channels displayed = 25 out of 31).