US 6587725

I'll search for authoritative information on this specific patent number, including any litigation or CAFC docket activity.

Let me check for any litigation or CAFC docket activity specific to this patent.

US 6,587,725 B1 — Summary

Identification check: All results retrieved correspond to the specific number 6587725 (US 6,587,725 B1). No results for similar numbers were substituted. Sources: Google Patents full text (the authoritative text supplied), plus Justia, FreePatentsOnline, and Unified Patents listings.

Bibliographic data

Abstract (verbatim structure)

A fully implanted functional electrical stimulator (20) and method for treating obstructive sleep apnea (OSA) that provides reliable detection/prediction of airway occlusion and relieves and/or prevents it by selective, direct electrical stimulation of the hypoglossal nerve (HG). The method/apparatus sense hypoglossal nerve electroneurogram (ENG) activity to detect or predict OSA, and the sensed nerve activity itself triggers the electrical stimulation to improve upper airway patency. An improved hypoglossal nerve electrode (10) interface (IC) permits simultaneous nerve-activity sensing and stimulation by eliminating stimulation artifacts that would otherwise trigger further erroneous stimulation.

Plain-language overview of the independent claims

There are only two independent claims — claim 1 (method) and claim 7 (apparatus). Claims 2–6 depend from claim 1; claims 8–12 depend directly or indirectly from claim 7.

• Claim 1 (method): A method of treating OSA in a human patient comprising (a) monitoring the patient for at least partial occlusion of the upper airways by sensing the electroneurogram activity of the patient's hypoglossal nerve, and (b) directly electrically stimulating that same hypoglossal nerve when at least partial occlusion occurs, as indicated by the sensed nerve activity. The core concept: the hypoglossal nerve's own ENG signal is the trigger for stimulating that nerve.

• Claim 7 (apparatus): An apparatus for treating OSA comprising (a) means for detecting at least partial upper-airway occlusion by sensing hypoglossal nerve ENG activity, and (b) means for directly electrically stimulating the hypoglossal nerve in response to that detected occlusion. This is drafted in means-plus-function form under 35 U.S.C. §112(f).

Notably narrower dependent features (not independent, but define the commercially salient subject matter): claim 4/10 require stimulus pulse trains with amplitude >0.2 mA; claims 5–6 and 11–12 require a cuff-type nerve electrode at least partially encircling the nerve, with sensing on a third contact positioned so that the impedance from it to each of the two stimulation contacts through the nerve tissue is substantially equal — the impedance-balancing arrangement that cancels stimulation artifact and EMG.

CAFC / litigation check

• I found no Federal Circuit (CAFC) docket, PTAB/IPR, or district-court case that names U.S. Patent 6,587,725 as an asserted patent. Searches returned nothing tying this specific number to the active Inspire v. Nyxoah disputes.
• The related 2025–2026 litigation between Inspire Medical Systems, Inc. and Nyxoah (Nyxoah SA) — e.g., D. Del. Case No. 1:25‑cv‑00667, and the consolidated proceeding referenced in Nyxoah's 2026 SEC filing — asserts different patents: U.S. 10,898,709; 11,806,526; 11,850,424 (Inspire‑asserted) and 8,700,183; 9,415,215; 9,415,216 (Nyxoah‑asserted). Patent 6,587,725 is not among them.
• Patent 6,587,725 is heavily cited by later hypoglossal-nerve-stimulation patents (Google Patents lists 100+ "Cited By" entries and Unified Patents lists over 300 "Referenced By" records), and its specification is cited as background art in later filings (e.g., EP 2 038 005 B1; the Chinese family member CN 101495178 B, which characterizes Durand et al. as only stimulating upon occlusion and therefore not preventing occlusion). That is citation activity, not litigation.

Uncertainty / limitations

• I do not have direct authenticated USPTO Patent Center or PACER/CAFC database access in this session; the bibliographic and legal-status data above come from Google Patents, Justia, FreePatentsOnline, Unified Patents, and secondary SEC/court-reporting pages. Legal status ("Expired – Lifetime") is the sources' representation, not a legal conclusion.
• I found no evidence of litigation or CAFC review of this patent; that is a negative finding from the searches performed and should not be read as proof that none exists.
• Minor date discrepancies appear across secondary sources (e.g., one Unified Patents entry lists a priority date of 1998‑07‑26, and "Filed: Mar 26, 2001" on Justia reflects national-stage entry rather than the 1999‑07‑27 international filing date). The patent's own text (priority 1998‑07‑27; PCT filed 1999‑07‑27) is authoritative where sources conflict.

I'll search litigation databases and related sources for this specific patent number.

Let me search litigation-specific databases and case-number formats.

Let me try more targeted queries, including case-caption patterns and the patent owner's name.

Litigation search: U.S. Patent No. 6,587,725

Bottom line: I found no record — in any of the sources searched — of litigation, PTAB/IPR proceeding, ITC investigation, or Federal Circuit appeal naming U.S. Patent 6,587,725 as an asserted or challenged patent. This is a negative finding from the searches performed, and carries the caveats noted at the end.

---

1. What was searched (and what the searches actually returned)

2. Important disambiguation — do NOT conflate with other activity

Two items surfaced that involve the digits "658"/"7725" but are different patents or different roles, and I flag them expressly per the rule on literal interpretation:

(a) The Dynapass "IPR2023-00367 / '658 patent" hits are NOT this patent. The PTAB petition served via ptacts.uspto.gov concerns U.S. Patent No. 6,993,658 ("DynaPass," payment-authentication/NPE litigation — Dynapass IP Holdings LLC v. Regions Financial, et al., 2:22‑cv‑00215 (E.D. Tex. June 17, 2022), and eight companion E.D. Tex. suits). That is a different number (6,993,658) and an unrelated technology. It is not US 6,587,725.

(b) The Inspire v. Nyxoah docket hits are NOT this patent. Documents from D. Del. (e.g., gov.uscourts.ded.89290, and the consolidated Inspire v. Nyxoah matter referenced in the earlier summary, incl. 1:25‑cv‑00667) assert the '709 (U.S. 10,898,709), '215, '216 patents — not '725.

(c) The IPRs against U.S. 10,898,709 (Inspire) treat Durand as prior art, not as the patent-in-suit. In the Loeb declaration filed in support of a petition for IPR of U.S. 10,898,709 (ptacts.uspto.gov/ptacts/public-informations/petitions/1558660, served Dec. 18, 2025), the prevailing invention is U.S. 10,898,709, and "Durand" is relied on as the primary prior-art reference (with Hoegh, Headley, Tran). The '725 patent is cited as prior art against another party's patent; it is not itself being litigated. That is citation activity, not litigation.

(d) Foreign citation only. WO 2005/087313 A1's ISR cites WO 2000/006249 A2 (Case Western Reserve Univ.) with its family member US 6,587,725; CN 101495178 B (Philips) cites US 6,587,725 as background and distinguishes it ("刺激...仅… methods... does not prevent occlusion"). EPO EP 2 038 005 B1 lists US 6587725 B, Durand under "References Cited in the Description." All are prosecution/opposition citations — not litigation.

3. Reasonable explanation for the absence

US 6,587,725 issued 2003‑07‑01, was assigned to Case Western Reserve University, and expired for failure to pay maintenance / reached end of term (anticipated expiration 2019‑07‑27; "Expired – Lifetime"). An academic-owned, long-since-expired patent is a low-probability candidate for a district-court assertion docket. The commercial activity in this space ran instead through the separate CWRU-continuation patents (e.g., U.S. 7,680,538; US 2006/0224211) and through the modern Inspire / Nyxoah / LivaNova / ImThera portfolios.

4. Caveats / limitations (stated explicitly rather than glossed)

• I do not have authenticated PACER or CAFC CM/ECF access in this session. The searches were against Unified Patents, Google Patents, Justia/FreePatentsOnline, PTAB (ptacts.uspto.gov), and Archive.org court-document mirrors.
• PACER dockets are not full-text indexed by patent number in the way patent databases are; a suit that was filed and dismissed early, or settled pre-service, could exist without a searchable patent-number link.
• Therefore: treat the above as "no known/recorded litigation found," not as an affirmative proof that none ever occurred.
• If you need a belt-and-suspenders confirmation, the check would be a Lex Machina / Docket Navigator full-text docket search on "6,587,725" (including terminals and complements), plus a PTAB API search for the patent number across all AIA trial types (IPR/PGR/CBM/derivation), which would be the definitive negative.

If forced to a single conclusion for this record: no litigation is known involving U.S. Patent 6,587,725.

I'll verify the structured data against live sources before drafting.

Let me run additional targeted checks before concluding.

One more check on older-window (2012–2019) filings and any PGR/CBM activity.

Proceedings overview

There are zero AIA trial proceedings on file for US 6,587,725 — the PTAB proceedings block supplied with this analysis (USPTO Open Data Portal, the canonical source) returns an empty set, and independent web searches across PTAB E2E, the USPTO Patent Trial and Appeal Board decisions pages, CourtListener/Federal Circuit reporting, and the litigation press surface no IPR, PGR, or CBM petition, no institution decision, no Final Written Decision, and no appeal naming this patent. The bottom-line defensive posture is therefore the opposite of a hardened patent and the opposite of a canceled one: it is an untested patent whose threat has been neutralized by time rather than by the Board. No claim of 6587725 has ever been adjudicated by the PTAB — but as of 2026-09-28 the patent has been expired for more than seven years (expiration 2019-07-27), so an IPR is likely unnecessary as a defensive tool and, in practical terms, there is no live damages theory (see Strategic summary).

---

No proceeding to summarize — and why the template is left empty

The task template calls for one ### {PROCEEDING_NUMBER} — {Petitioner} v. {Patent Owner} block per proceeding, ordered most-impactful first. There are no proceeding numbers to populate. I am deliberately not fabricating an IPR number, a petitioner, a panel, or a disposition. What follows is the record of what was checked and what was found.

Key technical reason no PGR or CBM can exist: US 6,587,725 is a pre-AIA patent (priority 1998-07-27, national stage of PCT/US1999/016913). It therefore has no claim with an effective filing date on or after 2013-03-16, which makes it categorically ineligible for Post-Grant Review. CBM review is unavailable as well — the claimed subject matter (implanted nerve stimulation) is not a "financial product or service" under § 18 of the AIA. Inter partes review under §§ 311–319 is the only AIA vehicle that could ever have reached this patent.

Minor contradiction to flag: the structured block and my searches agree there is no PTAB activity, but note that the block is a snapshot of an ingest pipeline. Pre-institution terminations (petitions dismissed before an institution decision and without a public Board paper) can be under-indexed in ODP. I found no secondary evidence of such a filing either, so the finding stands, but it is a negative finding from the sources checked, not proof of nonexistence.

---

Strategic summary

Cancelled vs. sustained vs. untested. This is the cleanest possible answer: all twelve claims — independent claims 1 and 7 plus dependents 2–6 and 8–12 — are UNTESTED. Not one claim has been canceled, and not one has been sustained, because no claim has ever been before the Board. There is no narrowing reissue, no certificate of correction canceling claims, and no disclaimer of record that the prior section surfaced. For a defendant, this cuts both ways: there is no claim-level roadmap from the PTAB telling you which claims are weak, and there is no estoppel binding anyone — including you. The validity of claims 1–12 remains exactly as it left the examiner's desk in 2003, with only two references of record.

Estoppel landscape. Section 315(e)(2) estoppel does not exist for this patent, because estoppel only attaches upon a Final Written Decision in a proceeding that reaches that stage. With no IPR ever instituted, no petitioner and no privy is barred from raising any § 102 or § 103 ground, and no district-court defendant faces a § 315(e)(2) trap. The entire prior-art universe is available to a defendant — including the dense later literature the patent itself helped generate. Google Patents and Unified Patents list hundreds of citing documents (Apnex/Cyberonics, Imthera, Nyxoah, Inspire, Philips' hypoglossal-nerve family including WO 2008/005903 → US 7,660,632, Cardiac Pacemakers, LivaNova, and others). Some of these post-date the 1998 priority date and are therefore not prior art against 6587725, but the sheer volume of the intervening hypoglossal-nerve literature (e.g., the Sahin/Durand/Haxhiu 1997–2000 journal publications, and the pre-1998 Meer and Testerman references already of record) means a § 102/§ 103 challenge would be built on a much broader evidentiary base than the one-reference-efficiency the examiner worked with. The absence of any IPR is itself the signal the task prompt describes: normally, a patent this widely cited in a commercially hot space (hypoglossal nerve stimulation for OSA) would have attracted at least one petition. That it did not is consistent with the patent having expired before the modern Inspire/Apnex/Imthera/Nyxoah commercial wave reached full intensity — there was little to defend against and, eventually, little damages exposure to attack.

Pattern signals. (a) No petitioner has ever filed against this patent, so there is no repeat-petitioner history to analyze. (b) The patent owner (Case Western Reserve University) has never had to appeal a PTAB decision on this patent, because there was none — contrast this with the aggressive PTAB/CAFC posture visible in the other OSA families (the LivaNova/Inspire US RE48024/RE48025 reissue line; the Inspire v. Nyxoah D. Del. 1:25-cv-00667 action asserting US 10,898,709, 11,806,526, 11,850,424 against Nyxoah's US 8,700,183, 9,415,215, 9,415,216). (c) No defensive aggregator is in the chain for 6587725. Unified Patents maintains a monitoring/prior-art page for the patent (https://portal.unifiedpatents.com/patents/patent/US-6587725-B1) but has not filed against it; that page is citation infrastructure, not a challenge.

The expiration point dominates everything. The patent's term ran from 1998-07-27 and anticipated expiration was 2019-07-27. Two consequences matter more than any hypothetical IPR:

1. No injunctive relief is available. With no ongoing infringement after expiration, a plaintiff cannot obtain an injunction, and eBay/ongoing-infringement theory evaporates.
2. The § 286 six-year damages lookback now reaches zero. As of today, 2026-09-28, a complaint filed now could recover damages only for infringement occurring on or after 2020-09-28 — a date after the patent expired. The entire recoverable window is time-barred. A suit asserting 6587725 today is, on its face, a suit with no damages remedy and no injunctive remedy.

---

Recommended next steps

If you are a defendant and receive a demand letter or complaint citing US 6,587,725:

• Lead with expiration, not with the PTAB. Because the structured record shows no AIA trial proceeding on file, there is no FWD to link to and no claim to quote as canceled. The dispositive facts are the 2019-07-27 expiration and the § 286 six-year bar, which as of 2026-09-28 leaves no compensable damages period. Confirm the expiration date against the USPTO Patent Center record (https://patentcenter.uspto.gov) and the fee/maintenance history — the patent's own text and Google Patents (https://patents.google.com/patent/[US6587725B1](/patent/US6587725B1)/en) both show "Expired – Lifetime."
• Assume the asserted claims are the narrow ones if the letter is specific. Independent claim 1 (method: sense hypoglossal ENG → directly stimulate that nerve on occlusion) and independent claim 7 (means-plus-function apparatus, § 112(f)) are the only independents. Claims 4/10 (pulse amplitude > 0.2 mA) and claims 5–6 / 11–12 (cuff electrode at least partially encircling the nerve, with a third sensing contact impedance-balanced to each stimulation contact) carry the commercially salient limitations. A § 112(f) indefiniteness/means-plus-function construction fight over claim 7 is available and untested.
• Free rein on prior art. With no IPR and therefore no § 315(e)(2) estoppel, you may raise any § 102/§ 103 ground in district court or before the Board. The two references of record (US 4,830,008 Meer; US 5,540,732 Testerman) are the starting point, not the ceiling.
• An IPR is legally available but strategically marginal. Expired patents are not categorically immune from IPR, but the Board weighs settled expectations and the absence of a live controversy, and you would be spending six figures to invalidate claims that already cannot support a damages award. Recommendation: raise expiration/§ 286 in a Rule 12 motion or a stay-and-dismiss posture rather than filing an IPR.
• Preserve the negative finding properly. If a plaintiff asserts this patent, note in your file that (i) the USPTO ODP AIA-trial dataset shows no PTAB proceedings for 6587725, and (ii) no district-court, ITC, or CAFC case names it as an asserted patent. If the plaintiff's demand letter implies PTAB history or validity adjudication, that implication is unsupported — but do not characterize the absence of PTAB activity as an admission; it is simply the state of the record.

If you are monitoring rather than defending: the practical takeaway is that 6587725 is best treated as prior art and a citation hub, not as a live assertion risk. Its value now is as § 102/§ 103 material against the later hypoglossal-nerve patents in the field — which is presumably why it appears in Unified Patents' monitoring index and why the Chinese family member CN 101495178 B distinguishes Durand et al. as stimulating only upon occlusion rather than preventing it (a characterization you can use either direction in claim-construction and validity disputes in the later families).

Caveat on sourcing: no PTAB proceeding, FWD, or CAFC opinion is cited above because none exists for this patent; accordingly there is no USPTO PTAB E2E paper link or CourtListener link to provide. All negative findings reflect the structured ODP block plus searches conducted on 2026-09-28, and should not be read as proof that no pre-institution termination ever occurred off-index.

I'll research the assignment chain for this patent using the available sources.

Let me dig deeper into the assignment records and any subsequent transfers.

Let me try to locate the specific USPTO assignment reel/frame records and confirm there are no post-issuance transfers.

Note on source availability

The USPTO Assignment Center / Patent Assignment Search could not be fully queried within this session (I reached my tool-call limit before retrieving the recorded reel/frame images). What follows is built on the Google Patents legal-events record for US 6,587,725 (the authoritative text supplied in the prompt), corroborated by the inventor biography material and secondary listings. I did not fabricate any reel/frame number. Where the reel/frame is unknown, I say so explicitly.

---

Inventors

Pattern note (not alarming): All three named inventors were CWRU‑affiliated at filing, which is exactly why the single recorded assignment runs from the three inventors to CWRU rather than to a third party. One inventor — Sahin — left CWRU for Louisiana Tech University in 2001, i.e. within ~3 years of filing, but that is the ordinary academic post‑doc‑to‑faculty move. His CWRU post‑doc ran 1998–2001 and his bio sketch lists the '725 patent as his PhD‑dissertation work product. There is no evidence of a coordinated inventor exodus preceding a portfolio sale — all three were still CWRU‑affiliated on the 2001 assignment date.

---

Original assignee

• Entity on the issued patent: Case Western Reserve University (front page "Current Assignee"; assignment to CWRU recorded 2001‑03‑26 per Google Patents legal events). At filing, the front page shows the original assignee as "Individual" (the inventors), which is why the 2001 documentation appears as a transfer.
• Primary line of business: Non‑profit research university (higher education + sponsored research; tech‑transfer via its Office of Technology Management). It is not a device manufacturer.
• Did it ship a product embodying the claims? No. CWRU did not commercialize an implantable hypoglossal‑nerve stimulator itself. The claimed subject matter (closed‑loop HG stimulation with an artifact‑cancelling tripolar cuff) was demonstrated only in the two chronically instrumented Beagles described in the specification (FIGS. 1–4), not in a commercial product.
• Current status: Operating. CWRU remains the assignee of record on the available sources; the patent is Expired – Lifetime (anticipated expiration 2019‑07‑27, 20 years from the 1999‑07‑27 PCT filing). No bankruptcy, dissolution, or acquisition of the assignee is associated with this patent.
• Family note: CWRU also owns the related, later‑issued US 7,680,538 ("Method of treating obstructive sleep apnea using electrical nerve stimulation," Durand & Yoo), confirming CWRU retained the hypoglossal‑stimulation IP rather than selling it to an NPE.

---

Assignment timeline

Chronological list of every recorded assignment located:

• Executed ~early 2001 (exact execution date not captured) / recorded 2001‑03‑26 — Reel unknown — not retrieved in this session
  • Conveyance: Assignment (inventor‑to‑employer / "ASSIGNMENT OF ASSIGNORS INTEREST, SEE DOCUMENT FOR DETAILS")
  • Assignor: SAHIN, MESUT; HAXHIU, MUSA A.; DURAND, DOMINIQUE (all three)
  • Assignee: CASE WESTERN RESERVE UNIVERSITY
  • Correspondent: not captured (the Google Patents event states only "ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS)" — the recording attorney/firm was not exposed in the retrieved record). Cannot confirm recurrence of a correspondent.
  • Context: Routine academic inventor‑to‑university assignment — the three inventors were CWRU students/faculty, and CWRU's IP policy captured the PCT‑filed invention. Not a commercial or NPE‑driven transfer.

No other recorded assignments exist. Google Patents' legal‑events stream for US 6,587,725 shows only: (1) 1999‑07‑27 application filed by "Individual"; (2) 1999‑07‑27 priority to US 09/744,640; (3) 2001‑03‑26 assigned to CASE WESTERN RESERVE UNIVERSITY; (4) 2003‑07‑01 grant; (5) 2003‑07‑01 publication; (6) 2019‑07‑27 anticipated expiration. There is no post‑issuance assignment, security agreement, merger, release, or license recording on the face of this record.

⚠️ Caveat: I could not open the Assignment Center reel/frame images for the 2001‑03‑26 entry, so the reel/frame and correspondent fields are unverified. Anyone relying on this should pull the record directly at the USPTO Assignment Center (search "6587725") to confirm the reel/frame and correspondent.

---

Timeline diagram

timeline
    title Ownership of US 6587725
    1998 : Filed by CWRU inventors
    1999 : PCT filed 27 Jul
    2001 : Assigned to Case Western Reserve Univ
    2003 : Patent issued 1 Jul
    2019 : Patent expires 27 Jul

---

NPE / troll-pattern signals

---

Verdict

Insufficient data (no records, or only the original assignment).

The only recorded transfer in the entire chain is the 2001‑03‑26 assignment from inventors Sahin, Haxhiu, and Durand to Case Western Reserve University (Google Patents legal events; reel/frame not retrieved this session). There are no post‑issuance assignments, no shell LLCs, no known asserter, no cascading transfers, and no defensive‑aggregator termination — every NPE signal above is either not present or "no data." CWRU is an operating non‑profit university assignee that never asserted this patent; it is not a product‑shipping company either, so "operating‑company assertion" does not fit. This patent reads as a clean academic patent that aged out to its 2019‑07‑27 expiration.

Verification link: https://assignmentcenter.uspto.gov/ (search "6587725") — pull the 2001‑03‑26 record to confirm reel/frame and correspondent, which I could not retrieve here.

---

Contradictions / flags for the record

• Filing‑date ambiguity persists from the prior section: Google Patents shows 1999‑07‑27 (PCT international filing carried into the §371 record) and Justia shows 2001‑03‑26 (national‑stage entry). This matters for the assignment timeline because the 2001‑03‑26 CWRU assignment is the same date as the Justia "filing date," consistent with the assignment being recorded around national‑stage entry. The patent's own text (provisional 1998‑07‑27; PCT filed 1999‑07‑27) remains authoritative.
• Reel/frame and correspondent are unconfirmed. I explicitly did not invent them. If a future pass retrieves them, the "repeat correspondent" signal should be re‑evaluated — though with a single record it can only ever be a non‑finding.

I'll start by locating the USPTO record for this exact patent number and then pull the two references cited on its face, plus confirm which claims they map to.

Prior-Art Analysis — U.S. Patent No. 6,587,725 B1

Scope note / method. The patent number was interpreted literally as 6587725; no similar number (e.g., 6,993,658 or 6,587,726) was substituted. The authoritative full text you supplied lists exactly two front-page patent citations ("Patent Citations (2)"): US 4,830,008 and US 5,540,732. I verified each against its own patent record (Google Patents full text, USPTO-hosted .pdf images, and uspto.report grant pages). No retrieved source contradicted the supplied text. I do not have authenticated USPTO Patent Center/PAIR access in this session, so the citation set is taken from the patent's own front page as reproduced at https://patents.google.com/patent/[US6587725B1](/patent/US6587725B1)/en and the USPTO-hosted grant record.

---

1. The complete cited-reference set for 6,587,725

Not prior art to '725 (flagged to avoid confusion):

• US 60/094,263 (provisional) and PCT/US1999/016913 → WO 2000/006249 A2 are '725's own priority chain, not references.
• The ~116–319 "Cited By" documents (Apnex/Cyberonics '442 family, Inspire, ImThera, Nyxoah, Philips, etc.) all post-date the 1998‑07‑27 priority date and therefore cannot anticipate '725.
• The "Families Citing this family" items (US 2007/0213782 A1; EP 3 527 178 B1; US 2013/0030257 A1) likewise post-date and cite '725 as background, not as anticipating art.
• Sahin et al., J. Appl. Physiol. 87(6):2197–2206 (1999) — the inventors' own canine ENG study, cited in later Apnex prosecution — published after the '725 priority date and is the inventors' own work; it is not a §102 reference against '725.

Legal-status framing. '725 claims priority to 1998‑07‑27 and has an international filing date of 1999‑07‑27 (national-stage entry 2001‑03‑26). Both cited references issued more than one year before the earliest date, so each is prior art under pre-AIA 35 U.S.C. §102(a) and §102(b) (and independently under §102(e) as of its US filing date).

---

2. Reference 1 — US 4,830,008 (Meer)

Full citation. U.S. Patent No. 4,830,008, Method and system for treatment of sleep apnea, inventor Jeffrey A. Meer (Birmingham, MI). Application No. 07/156,186, filed 1988‑02‑16, as a continuation-in-part of App. No. 07/041,973 filed 1987‑04‑24; issued 1989‑05‑16. 23 claims. Primary Examiner Francis Jaworski; Attorney Brooks & Kushman. US class 607/42; CPC A61N1/3601. Sources: https://patents.google.com/patent/[US4830008A](/patent/US4830008A)/en ; https://uspto.report/patent/grant/4830008 ; https://insight.rpxcorp.com/patent/US4830008A

Brief description. A closed-loop sleep-apnea system: a monitor (14) detects inspiratory effort (diaphragm EMG, phrenic-nerve action potentials, intrathoracic pressure via a pressure transducer, or impedance pneumogram) and an implanted signal generator (16) applies stimulation to the nerves activating the upper-airway dilator muscles — geniohyoid, genioglossus, digastric, stylopharyngei, mylohyoid — to hold the airway patent. The effector electrode (32) can be placed on the motor nerve of those muscles (i.e., the hypoglossal nerve for the genioglossus; Fig. 5 shows the analogous phrenic arrangement) and, per Figs. 2 and claim 16, that same effector lead can double as a sensor of action potentials in those nerves. In the preferred embodiment, sensed normal upper-airway muscle activity inhibits stimulation (claim 23: "sensing upper airway muscle activity and inhibiting the release of electrical signals if normal upper airway activity is detected").

§102 anticipation mapping (element-by-element, against the issued claims of '725):

Closest §102 argument that a challenger could attempt. Meer claim 13 ("said monitor (14) senses action potentials in those nerves which innervate those muscles which maintain upper airway patency") and claim 16 (effector lead doubles as an action-potential sensor) could be pressed as disclosing nerve-action-potential sensing on the nerve that innervates the genioglossus — i.e., the hypoglossal nerve — followed by stimulation of that same nerve. That is the strongest anticipation theory against claim 1. It fails because Meer ties that sensing to inspiratory-effort monitoring and inhibitory control ("if normal activity is sensed, the release of electrical signals is inhibited"), not to detecting at least partial airway occlusion. So the limitation "monitoring … for at least partial occlusion … by sensing electroneurogram activity of a hypoglossal nerve" is not disclosed.

---

3. Reference 2 — US 5,540,732 (Testerman)

Full citation. U.S. Patent No. 5,540,732, Method and apparatus for impedance detecting and treating obstructive airway disorders, inventor Roy L. Testerman (New Hope, MN), assignee Medtronic, Inc. (Minneapolis, MN). Application No. 08/310,120, filed 1994‑09‑21; issued 1996‑07‑30. 12 claims. EP counterpart EP 0 702 977 A2/A3 (A2 published 1996‑03‑27). Google Patents currently lists the assignee as Inspire Medical Systems, Inc. (portfolio acquisition). US class 607/42. Sources: https://patents.google.com/patent/US5540732 ; https://uspto.report/patent/grant/5540732 ; https://insight.rpxcorp.com/patent/[US5540732A](/patent/US5540732A)

Brief description. An implantable pulse generator (IPG) with an impedance-sensing circuit in a common case senses transthoracic impedance to identify the onset of inspiration; the IPG then delivers inspiration-synchronous stimulation via a lead to a cuff electrode around the hypoglossal nerve (explicitly: "the genioglossus muscle which may be stimulated by a cuff electrode placed around the hypoglossal nerve"; "stimulation … to the nerve electrode 73 around the hypoglossal nerve 74"; a Medtronic Model 3990 Half Cuff Nerve Electrode). Detection is threshold/slope-based on the impedance waveform; stimuli are applied "at a level which restores patency in the airway."

§102 anticipation mapping:

Related same-family art (same inventor/assignee, same filing date) — useful §103 context, not on '725's face: US 5,540,731 (pressure detecting), US 5,540,733 (detecting and treating OSA), US 5,522,862, US 5,591,216, US 5,483,969, and US 5,344,438 (Testerman cuff electrode).

---

4. Bottom line

1. Neither cited reference anticipates any of claims 1–12 of US 6,587,725. Both are prior art under pre-AIA §102(a)/(b) (issued 1989 and 1996, well before the 1998‑07‑27 priority date), but each fails the core two-limb limitation of independent claims 1 and 7: (i) sensing the hypoglossal nerve's own electroneurogram and (ii) using that signal as the indicator of upper-airway occlusion that triggers stimulation of that same nerve. Meer senses inspiratory effort (and upper-airway muscle activity, used inhibitorily); Testerman senses transthoracic impedance to find inspiratory onset.

2. Closest §102 candidate: Meer (US 4,830,008) against claims 1, 3, 7, 9 — via its nerve-action-potential sensing on the nerve innervating the genioglossus (claims 13/16) and stimulation of that same nerve. It survives only because the sensing is tied to inspiratory-effort monitoring and inhibitory control, not to occlusion detection.

3. Best §102 candidate for the electrode-dependent claims: Testerman (US 5,540,732) against the electrode limitations of claims 5 and 11 (half-cuff around the hypoglossal nerve, stimulation through nerve tissue) — but not against claims 5/11 as a whole, since those depend from claims 1/7.

4. Claims 6 and 12 are not touched by either reference. The impedance-balanced third (sensing) contact — the tri-polar cuff that cancels stimulation artifact and common-mode EMG — is the point of novelty that distinguishes '725 over its own cited art, and it is the feature the specification (Fig. 6 / interface circuit IC) is built around.

5. Practical read of the prosecution record: with only two §102(b) references and both lacking the "hypoglossal ENG as the occlusion trigger" limitation, the cited art functioned as **§103* background rather than an anticipation set. That is consistent with later third-party characterizations of '725 as disclosing closed-loop, ENG-triggered hypoglossal stimulation (e.g., EP 2 038 005 B1 lists "US 6587725 B, Durand" and US 4,830,008 / US 5,211,173 as description-level background, and CN 101495178 B distinguishes Durand et al. as stimulating upon occlusion).

---

Caveats

• The reference set above is the front-page examiner citation set as reproduced in the supplied authoritative text and corroborated by the Google Patents / uspto.report grant records. I did not open a primary USPTO Patent Center/PAIR image of the paper front page in this session; if a 1449/IDS item or an examiner's "NPL" citation exists on the physical front page, it would not appear in the patents.google.com structured citation counts, so a Patent Center pull remains the belt-and-suspenders check.
• "Anticipation" above is analyzed against the issued claims as printed and is a technical mapping, not a legal conclusion. The Meer claim‑13/16 argument in §2 is the one plausible §102 theory I could construct; it should be treated as an argument a challenger could make, not as a disposition of claim validity.
• Filing/priority dates are stated per each reference's own record; where secondary databases differ, the reference's own front page controls.

I'll ground the §103 analysis in the actual cited prior art first.

Let me pull the closely-related Testerman/Medtronic family art and the key non-patent literature that predates the 1998 priority date.

Obviousness Analysis — U.S. Patent No. 6,587,725 B1 under 35 U.S.C. § 103

Prepared: 2026-04-26 · Patent-in-analysis: US 6,587,725 B1 (Durand, Sahin, Haxhiu; Case Western Reserve University)
Source of record (authoritative): https://patents.google.com/patent/US6587725/en
Analytical posture: Hypothetical/defensive validity assessment (not a litigation position for any party). Builds on — and does not repeat — the bibliographic and litigation findings already generated. No litigation is known involving this patent, so this §103 analysis is a fortiori academic; it is nonetheless worth doing because '725 is the ancestor reference cited against later, live patents (e.g., U.S. 10,898,709).

---

I. Governing law and the effective date

The '725 application claims priority to U.S. provisional 60/094,263 (filed 1998‑07‑27) via PCT/US1999/016913 (filed 1999‑07‑27), and issued 2003‑07‑01. All critical dates precede the AIA's March 16, 2013 first-to-file change; therefore pre‑AIA § 103(a) applies and prior-art status is judged under pre‑AIA § 102(a)/(b)/(e)/(f)/(g).

Obviousness is assessed under the Graham v. John Deere framework (scope/content of the prior art; differences between the prior art and the claims; level of ordinary skill; secondary considerations), as reanimated by KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007). Critically for this patent, KSR authorizes reliance on: (i) combining known elements per known methods to yield predictable results; (ii) simple substitution of a known element to obtain a predictable result; (iii) use of a known technique to improve a similar device in the same way; (iv) a "finite number of identified, predictable solutions"; and (v) design incentives/market forces.

Level of ordinary skill (PHOSITA), 1998: a biomedical engineer or a clinician-engineer with 2–4 years in implantable neurostimulation, functionally literate in (a) nerve-cuff electrode design, (b) biopotential recording/artifact rejection, and (c) respiratory neurophysiology of the upper airway. This is a high skill level — the field was already populated by implantable IPGs, perineural cuff electrodes, and commercial programmable stimulators (Medtronic ITREL generation).

---

II. What each claim requires (element decomposition)

Independent claim 1 (method) and claim 7 (apparatus, means-plus-function under §112(f)) require, in substance:

Dependent features: cl. 2/8 = occlusion identified when ENG rises above a baseline level by a select amount; cl. 3/9 = amplify the sensed ENG and use the amplified signal as trigger; cl. 4/10 = stimulus pulse train amplitude > 0.2 mA; cl. 5/11 = nerve electrode at least partially encircling the HG, pulses passed between first and second contacts through nerve tissue; cl. 6/12 = third (sensing) contact arranged so the impedance from it to each of the two stimulation contacts through the nerve is substantially equal.

The '725 specification is unusually candid about how much of the problem space was already known. Its own Background admits: (1) it was "generally known that the flow of inspired air is doubled by stimulation of the main branch of the hypoglossal nerve"; (2) medial-branch stimulation "was nearly as effective"; (3) "Attempts have been made to improve upper airway patency in humans during sleep via direct electrical stimulation of the hypoglossal nerve"; and (4) synchronizing that stimulation to respiration was already done using hypopharyngeal/esophageal pressure, nasal/mouth thermistor airflow, and tracheal inter-ring strain gauges. These are applicant-admitted prior art (AAPA) and are fair game in any §103 combination.

That concession narrows the case dramatically: the only thing '725 adds over the admitted art is the identity of the sensor — the HG's own ENG rather than pressure, airflow, strain, or muscle EMG.

---

III. The prior art of record and the immediately adjacent art

A. Art cited on the face of the patent ("Patent Citations (2)")

1. US 4,830,008 — Meer, "Method and system for treatment of sleep apnea" (issued 1989‑05‑16).
https://patents.google.com/patent/[US4830008A](/patent/US4830008A)/en · https://uspto.report/patent/grant/[4830008](/patent/4830008)

Meer is the primary reference and is closer than its age suggests. It discloses an implantable monitor (14) + implantable electrical signal generator (16) + effector lead (26)/electrode (32); monitoring of inspiratory effort; generating electrical signals "in response to" the monitoring; and applying those signals "to those nerves which activate the patient's upper airway muscles" (Meer cl. 1). Specific additional disclosures that map onto the '725 dependent claims:

• Meer cl. 13: "said monitor (14) senses action potentials in those nerves which innervate those muscles which maintain upper airway patency." → Neural (ENG) sensing in the nerve supplying the upper-airway dilators. The hypoglossal nerve is the nerve innervating the genioglossus/geniohyoid — Meer cl. 17–18 expressly name the genioglossus and tongue-protrusion effect that HG stimulation produces (a fact '725 also admits).
• Meer cl. 16 / FIG. 2: the effector lead is also connected to the monitor to sense action potentials at the effector electrode (32) in those nerves innervating the upper-airway muscles — i.e., one electrode performing both stimulation and neural sensing on the same nerve, which is the structural premise of '725's claims 5/6 and 11/12.
• Meer cl. 7: monitoring includes "comparing the contraction … to a predetermined threshold contraction" — threshold-based triggering.
• Meer cl. 2/23: sensing upper-airway muscle activity and inhibiting release of stimulation if normal activity is detected — a closed loop in which sensed bioelectric activity gates stimulation.
• Meer cl. 21–22: the generator is programmable as to voltage, current, pulse width and frequency.
• Meer, FIG. 5: sensor electrode placed around the cervical portion of the phrenic nerve to detect action potentials — a perineural cuff used purely for sensing nerve action potentials on a cervical nerve, i.e., squarely a nerve-ENG sensor.

2. US 5,540,732 — Testerman, "Method and apparatus for impedance detecting and treating obstructive airway disorders" (Medtronic, Inc.; issued 1996‑07‑30).
https://patents.google.com/patent/US5540732 · https://uspto.report/patent/grant/[5540732](/patent/5540732)

Testerman is, for §103 purposes, the more dangerous reference because it supplies nearly every remaining element:

• Fully implantable IPG (Medtronic ITREL II 7424 modified) — https://uspto.report/patent/grant/5540732
• Detection of obstructive events: Testerman FIGS. 3, 4a–4b expressly show the respiratory-effort waveform changing "at the onset of obstructive apnea," with inspiratory peaks becoming significantly greater in amplitude — i.e., detecting the onset of upper-airway obstruction from an implanted sensor.
• Stimulation of the hypoglossal nerve: Testerman's FIG. 5 and specification recite "a cuff electrode placed around the hypoglossal nerve" and "the pulse generator 60 through the lead 72 to the nerve electrode 73 around the hypoglossal nerve 74," with an expressly identified Medtronic Model 3990 Half Cuff Nerve Electrode.
• Closed-loop, threshold-above-baseline triggering (Testerman cl. 4–5): "determining a baseline average value for the amplitude of the monitored signal; determining a threshold value as a function of the baseline average value"; "When the threshold voltage is achieved by the signal from the sensor, stimulation is enabled." This is a near-verbatim anticipation of '725 claims 2 and 8 (save for the sensor modality).
• Amplification/conditioning chain (Testerman FIG. 7): impedance signal 89 → amplifier/filter 87 (which "filters artifacts from the signal") → automatic gain control 88 → A/D 91 → microprocessor 100 → stimulus shaper 103. This reads on '725 claims 3 and 9.
• Stimulus "at a level which restores patency in the airway" and physician-set programmable amplitude → the '725 claim 4/10 stimulus-amplitude feature.
• Express motivation to abandon Meer's sensing approach: Testerman states that Meer's solution ("monitor the action potentials of the upper airway muscles to determine when the patient is awake and to commence stimulation only when normal upper airway muscle activity is not detected") "presents many practical difficulties in implementation and can lead to inappropriate stimulation or failure to stimulate reliably." A §103 motivation can be supplied by the prior art's own criticism — and here the criticism is explicit and directed at the sensing/triggering sub-problem the '725 patent claims to solve.

B. The Testerman/Medtronic family and the contemporaneous art cluster (available as §102(a)/(b) references)

• US 5,591,216 (Testerman et al., Medtronic, 1997‑01‑07) — "Method for treatment of sleep apnea by electrical stimulation" (cited via https://patents.google.com/patent/WO2015123360A1/en#22).
• US 5,540,731, US 5,540,733, US 5,522,862, US 5,483,969 — same Medtronic/Testerman family (pressure-, impedance- and respiratory-waveform-based OSA stimulation).
• US 5,174,287 (Kallok) — sensors monitor diaphragm EMG and thoracic and upper-airway pressures; "whenever … the pressure sensors show an abnormal pressure differential across the airway, the presence of obstructive sleep apnea is assumed and electrical stimulation is applied to the musculature of the upper airway." (Recited in Testerman '732's background at the URL above.) This is multi-signal occlusion detection → stimulation.
• US 5,211,173 / US 5,215,082 (Kallok et al.) — sensors measure effectiveness of the stimulation, and stimulus amplitude/pulse width or intensity are modified in response — i.e., a closed loop on the therapy itself.
• US 5,123,425 (Shannon) / US 5,178,156 (Wataru) — apnea event detection triggering upper-airway/genioglossus stimulation.
• US 5,190,053 (Meer) — intra-oral sublingual genioglossus stimulation electrode.

C. Non-patent literature predating the 1998‑07‑27 priority date

• Goding GS Jr, Eisele DW, Testerman R, Smith PL, Roertgen K, Schwartz AR. "Relief of upper airway obstruction with hypoglossal nerve stimulation in the canine." Laryngoscope 1998;108(2):162–169 (published Feb. 1998, DOI 10.1097/00005537-199802000-00003). https://pubmed.ncbi.nlm.nih.gov/[9473063](/patent/9473063)/
  This is the single most probative non-patent reference. It discloses a tripolar half-cuff electrode (Medtronic Model 3990A) placed around the medial branch of each hypoglossal nerve — and describes the electrode's architecture in terms that matter for claims 6/12: "The half-cuff configuration contains three platinum iridium electrodes, two outer anodes (guards), and one central cathode. The electrode cuff is 3 mm in diameter and 9 mm long." Full text available as an exhibit at https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1558660](/patent/1558660)/download-documents?artifactId=zIDhqIEJCV-G5SVaCyoIInC0zbfalmaGr9Djot66uP-q78am6pvioNE (served 2025‑12‑18, as prior art against U.S. 10,898,709 — not as a challenge to '725). Goding further reports chronic stimulation (30 Hz, 90 µs, 50% duty cycle) with no nerve damage, and that "This study provides support for clinical trials of hypoglossal stimulation for obstructive sleep apnea."
• Schwartz AR, Eisele DW, Hari A, Testerman R, Erickson D, Smith PL. "Electrical stimulation of the lingual musculature in obstructive sleep apnea." J. Appl. Physiol. 1996;81(2). https://www.semanticscholar.org/paper/Electrical-stimulation-of-the-lingual-musculature-Schwartz-Eisele/a79cdc161d97f982e5ebc4c2ba0bff5413790b63 — stimulation of lingual muscles can increase or decrease airflow depending on the muscle stimulated, without arousing the patient from sleep.
• Smith PL, Eisele DW, Podszus T, Penzel T, Grote L, Peter JH, Schwartz AR. "Electrical stimulation of upper airway musculature." Sleep 1996;19(10 Suppl):S284–S288 — ES of the hypoglossal nerve reduces upper-airway collapsibility and increases flow in animals and humans, with no associated cortical arousal; a half-cuff electrode on the nerve produced consistent tongue protrusion and improved airflow.
• Eisele DW, Smith PL, Alam DS, Schwartz AR. "Direct hypoglossal nerve stimulation in obstructive sleep apnea." Arch. Otolaryngol. Head Neck Surg. 1997;123(1):57–61. https://pure.johnshopkins.edu/en/persons/david-eisele/publications/

Collectively, the 1996–1998 literature establishes that, months before the '725 priority date, HG cuff stimulation in humans for OSA was not merely suggested but performed, and that the electrode geometry of choice was a tripolar cuff with a central contact flanked by outer contacts.

---

IV. Element-by-element mapping

Claim 1 / claim 7 (independent)

Assessment: Claim 1 is not literally anticipated — no single reference senses HG ENG and uses it as the occlusion trigger. But every limitation is disclosed, and the sole residual difference is the selection and location of the sensor element. Under KSR that is a textbook substitution candidate.

Claim 2 / claim 8 (baseline + "select amount")

Testerman '732 cl. 5 is a near-identical recitation in a different modality: "determining a baseline average value for the amplitude of the monitored signal; determining a threshold value as a function of the baseline average value," plus cl. 4 ("detecting a threshold value"). Meer '008 cl. 7 discloses comparison "to a predetermined threshold." Swapping the parameter being baselined (impedance → ENG) is a substitution of one known measured variable for another with predictable effect. Strongly obvious.

Claim 3 / claim 9 (amplify the sensed ENG; trigger on the amplified signal)

Testerman '732, FIG. 7: amplifier/filter 87 → AGC 88 → A/D 91 → microprocessor 100 → stimulus shaper 103. The '725 amplifier 36 / band-pass 38 / rectifier-averager 40 / threshold detector 42 chain is the same architecture, with a different front-end. Very strongly obvious.

Claim 4 / claim 10 (stimulus amplitude > 0.2 mA)

This is a bare numeric range on a result-effective variable (stimulus current) within an admittedly programmable generator (Meer '008 cl. 22; Testerman's physician-set "peak amplitude 117 … at a level required by the patient"). Under In re Aller / In re Boesch, optimizing a result-effective parameter in a disclosed adjustable range is routine. There is no evidence of criticality: the '725 specification's own FIG. 4 data report that 0.2 mA "did not result in significant large changes in the pressure measurements" while 0.4 mA reversed the loading effect and 0.6 mA caused no arousal — i.e., the patent's data show a smooth, non-critical dose–response, with the entire effective range lying above 0.2 mA and the animal tolerating 50% more than needed. Obvious as routine optimization; the specification undercuts any criticality argument.

Claim 5 / claim 11 (cuff electrode at least partially encircling the HG; pulses between first and second contacts through nerve tissue)

Goding 1998, verbatim: "A tripolar cuff electrode (Model 3990A, Medtronic Inc.) was placed around the medial branch of each hypoglossal nerve"; perineural field stimulation between the outer anode (guard) contacts in a half-cuff that is "3 mm in diameter and 9 mm long." Testerman '732 discloses the Medtronic 3990 Half Cuff Nerve Electrode around the HG. Meer '008 FIG. 5 discloses a perineural cuff. Directly disclosed in combination; obvious.

Claim 6 / claim 12 (third sensing contact, impedance to each stimulation contact substantially equal)

This is the only limitation with a colorable nonobviousness story, and it must be assessed carefully:

• Structure is disclosed: Goding 1998's Model 3990A is literally a three-contact cuff — two outer guard anodes flanking a central cathode — "two outer anodes (guards), and one central cathode." In a physically symmetric tripolar cuff, the axial distance (and hence the tissue-path impedance) from the central contact to each outer contact is, by construction, substantially equal. The claim-6 limitation is therefore met by the geometry of a reference electrode already implanted on the human hypoglossal nerve, before reconsidering any circuitry.
• Function is a known technique applied predictably: using a matched-impedance, symmetric sense path to obtain common-mode rejection of a large stimulus artifact is standard biopotential-amplifier practice (differential recording, balanced/center-tapped inputs, bridge nulling). Testerman '732's own FIG. 7 recites an "amplifier/filter 87 which filters artifacts from the signal 89," and notes that "Additional filtering of the impedance signal may also be required in order to remove physiological artifacts" — i.e., the field was actively engaged with stimulus/physiologic artifact rejection on the very same sensing channel.
• Alternative solutions were known and few: the '725 patent itself, in its FIG. 1 animal work, used simple time-domain blanking ("disabling the output of the trigger enabling circuit 24 for approximately an inter-breath interval") — and its own specification concedes the cost of that approach ("the phasic bursts of the activity signal HGS' were completely obscured by the stimulation artifacts"). KSR's "finite number of identified, predictable solutions" rationale applies squarely: a PHOSITA facing stimulation artifact on a sense electrode placed on the stimulated nerve would predictably consider (a) blanking, (b) high-order filtering, (c) differential/balanced recording, and (d) relocating the sense contact to a low-artifact node — and the '725 patent's own disclosure presents (a) and (c) as the two options.
• Claim breadth: claim 6 is drafted as a result-based functional limitation ("substantially equal … impedance"), not as a recitation of the transformer TR1 / matched coils L1=L2 / trimming potentiometer R1 bridge of FIG. 6. On its face it reads on any balanced-geometry tripolar cuff, including Goding's.

Assessment: moderately strong obviousness case, but the weakest of the set. If the patentee were defending this claim, the argument would be that a balanced bridge (transformer-coupled push-pull drive with trimming) as a deliberate artifact-cancellation architecture — as opposed to mere blanking — was a specific design insight not suggested by cuff geometry alone. That is a real, if narrow, argument. But it is squarely the kind of "known technique to improve a similar device in the same way" that KSR forecloses.

---

V. Specific proposed combinations and the motivation to combine

Combination 1 — Meer '008 + Testerman '732 (attacks claims 1–3, 7–9, and provides the frame for the rest)

Rationale: Both references are in the same field (implantable electrical stimulation for sleep apnea), address the same problem (maintaining upper-airway patency during sleep), and use the same class of hardware (subcutaneously implanted generator + lead + nerve/muscle electrode). Meer supplies the nerve-action-potential sensing with threshold gating and single-electrode sense/stimulate concept; Testerman supplies the express hypoglossal-nerve cuff target, the fully implantable synchronous-stimulation architecture, and the baseline-plus-threshold detection logic.

The express motivation is supplied by Testerman itself: it identifies Meer's upper-airway-muscle-action-potential sensing as producing "many practical difficulties in implementation" and as capable of "inappropriate stimulation or failure to stimulate reliably," and it therefore proposes a different sensor. A PHOSITA following that express lead — "find a more reliable sensor for detecting the onset of obstruction" — has a finite menu drawn from the art (intrathoracic pressure, airflow, strain gauge, EMG, and nerve ENG) and, having already been told by Meer that the nerve supplying the target muscle can be sensed (Meer cl. 13, 16), would predictably place the sensor on the same nerve that the Testerman cuff is already encircling. That yields claim 1 with no change in principle of operation, merely in the choice of measured signal.

Combination 2 — Meer '008 + Testerman '732 + Goding 1998 (and/or Schwartz 1996, Smith 1996, Eisele 1997) (attacks claims 5–6 and 11–12)

Rationale: Goding, Schwartz, Smith and Eisele are all by overlapping author groups in the same institution/industry nexus (Testerman and Schwartz appear in both the patent and the literature; the Goding work was funded by Medtronic, givehttps://onlinelibrary.wiley.com/doi/10.1097/00005537-199802000-00003). The combination is a same-field, same-problem, same-physiology combination with strong KSR "common design lineage" motivation. Goding supplies: (i) the cuff electrode at least partially encircling the HG (claims 5/11), and (ii) the three-contact, symmetric outer-guard/central arrangement (claims 6/12). Testerman supplies the lead, IPG and stimulation parameters; Meer supplies the sensing/gating concept.

Combination 3 — Testerman '732 + Kallok '287 + Meer '008 (attacks claims 1–3, 7–9 without relying on Testerman's criticism alone)

Rationale: Kallok '287 discloses multi-parameter occlusion detection (diaphragm EMG and trans-airway pressure differential) leading to upper-airway stimulation, and Kallok '173/'082 disclose modifying stimulation based on sensed results. Meer supplies neural sensing and the single-electrode sense/stimulate structure. Testerman supplies the hypoglossal cuff and the implantable architecture. Motivation: consolidating the sensing function onto the stimulation lead itself, to reduce lead count, tunneled hardware, and implant complexity in a fully implantable system — an express design goal of '725 and a well-recognized incentive in implantable device design (the '725 patent lists "no external devices or percutaneous electrical leads" and "a single, multiple-contact electrode … rather than multiple, separate electrodes/sensors" among its own advantages).

Combination 4 — Applicant-Admitted Prior Art (AAPA) + Meer '008 (the cleanest §103 route to claim 1)

The '725 Background admits that (a) direct HG stimulation in humans for OSA had already been attempted; (b) medial-branch stimulation was nearly as effective; (c) the doubling of inspired airflow by HG stimulation was "generally known"; and (d) the art already synchronized that stimulation using pressure, airflow and strain-gauge sensors. With those admissions, the claimed method differs only in the choice of trigger signal. That choice is (i) disclosed in kind by Meer (nerve action potentials), and (ii) one of a small, enumerated set of obvious sensing modalities the art had already canvassed. As In re Peterson/KSR teach, selecting one of a few known options for a known purpose, with predictable results, is obvious.

---

VI. KSR rationales stated expressly

1. Combination of known elements, predictable result. Implantable generator + nerve cuff + physiologic sensor + threshold logic was the recognized template (Meer, Testerman, Kallok, Shannon, Wataru). Substituting the sensed variable changes no element of structure or principle of operation.
2. Simple substitution of one known element for another. HG ENG replaces transthoracic impedance (Testerman), intrathoracic pressure (Meer FIG. 3), airflow (Wataru), or muscle EMG (Meer cl. 2) as the trigger input. Predictable result: a trigger signal from the same nerve being stimulated, available earlier and with higher specificity to the target effector — a result '725 itself reports (HG ENG onset preceding esophageal pressure, albeit with a very large dispersion: 17 ± 196 ms, 220 breaths / 20 maneuvers — a standard deviation larger than the mean, which undercuts any "unexpected" early-detection advantage).
3. Known technique improving a similar device in the same way. Balanced/differential artifact rejection in a symmetric tripolar cuff, plus programmable stimulus amplitude.
4. Finite number of identified, predictable solutions. Five sensing modalities were on the table in 1998; the inventors picked one.
5. Design incentives / market forces. '725 lists full implantability and single-electrode redundancy as its own objectives — these are the classic engineering incentives to put the sensor where the electrode already is.
6. "Obvious to try." Given Testerman's explicit finding that the competing sensor approach was unreliable, a PHOSITA had reason to pursue the neural sensing route with a reasonable expectation of success — especially since Goding had already demonstrated that a cuff on the HG can both stimulate and be interrogated, and had shown that the expected failure mode (nerve damage from chronic stimulation) did not occur.

---

VII. The strongest counterarguments (and why they likely fail)

A. "Meer teaches away." Meer's preferred embodiment uses sensed upper-airway activity to inhibit stimulation, not to trigger it (Meer cl. 2, 23). A patentee could argue this signals the opposite of the '725 approach.
Rebuttal: Meer's own claim 7 and FIG. 6 box 42 compare sensed activity to a predetermined threshold and generate signals "at the appropriate instant during the respiratory cycle"; Meer cl. 13 senses nerve action potentials; and Meer's FIG. 5 uses a cervical nerve cuff purely as a sensor. Meer thus discloses the same signal used in both the trigger and inhibition roles. Teaching away requires a disclosure that "criticizes, discredits, or otherwise discourages" the claimed approach (In re Fulton); identifying reliability concerns in one sensing modality is a motivation to improve, not a teaching away from neural sensing.

B. "No reference discloses sensing hypoglossal ENG." True as a literal matter, and this is the patent's real foothold.
Rebuttal: Claims 1/7 do not require a particular recording technique, electrode count, or fiber population; and Meer's nerve-sensing claim is generic to "those nerves which innervate those muscles which maintain upper airway patency," of which the HG is the principal and expressly-named example (genioglossus). Under KSR, reciting a known species (HG) from a disclosed genus (nerves to the upper-airway dilators) for the purpose already disclosed is an obvious narrowing.

C. "The interface circuit (FIG. 6) is a specific, non-suggested architecture." This is the best argument, available only for claims 6/12 — and only if the patentee can persuade the tribunal to import the transformer-bridge implementation into a limitation that reads, on its face, to any impedance-balanced three-contact arrangement. The specification itself makes that difficult: claim 6 recites only "a first impedance … substantially equal to a second impedance," with no recitation of coils L1/L2/L3, turns ratio n1=n2, or potentiometer R1.

D. "Unexpected results." The specification's quantitative results (SNR 2.37 ± 0.74; correlation with AreaPes R = 0.82/0.88; 2–3× AreaPes reduction reversed to control levels) are confirmatory of a known therapy effect, not unexpected. Goding had already shown 0.1 → 1.6 L/s peak flow improvement with HG stimulation; Testerman had already claimed "stimulation applied at a level which restores patency in the airway." The only genuinely new datum — that HG ENG leads esophageal pressure — has a standard deviation (196 ms) exceeding its mean (17 ms) and therefore cannot support a robust "unexpected early prediction" argument.

E. "Long-felt need / failure of others." The need was recognized (Testerman says so expressly), which supports motivation to combine as much as it supports nonobviousness. And the commercial record is weak for this patent specifically: it is CWRU-owned and Expired – Lifetime (anticipated expiration 2019‑07‑27); the modern HGNS franchise runs through other CWRU continuations (U.S. 7,680,538; US 2006/0224211) and through third-party portfolios. There is no nexus between the '725 claims and any commercial product's success.

---

VIII. Claim-by-claim conclusion

Overall: Claims 2–5, 8–11 are highly vulnerable; claim 1/7 are vulnerable but contestable on the "sense the very nerve you stimulate" selection; claims 6/12 are the only claims with a genuinely arguable nonobviousness position, and even those are more likely than not to fall under KSR, particularly given Goding's symmetric three-contact cuff and the ordinary practitioner's familiarity with balanced/biopotential artifact rejection.

Expected lifecycle view: had this claim set been tested, the probable outcome is a §103 rejection of claims 1–5 and 7–11 on Meer + Testerman (+ Goding), with claims 6/12 surviving only if the patentee successfully narrowed them to the FIG. 6 transformer-bridge architecture — which would require importing limitations the claims do not recite.

---

IX. Caveats, uncertainties, and flagged contradictions

1. Contradiction with a previously generated section — flagged. The earlier "Patent summary" states the cited prior art includes "US 5,540,732 (Testerman, later assigned to Inspire Medical Systems)." The search results retrieved here show US 5,540,732 as assigned to Medtronic, Inc. (inventor Roy L. Testerman, New Hope, MN; application 08/310,120; Medtronic appears throughout the specification as the IPG supplier and as the model-vendor: https://uspto.report/patent/grant/5540732 and https://patents.google.com/patent/US5540732). I found no evidence of an assignment of the '732 patent to Inspire Medical Systems; the Inspire-related hits in the search results were citations to '732 in later Inspire/Apnex/LivaNova documents and a family member (US 5,591,216) — not chain-of-title. Treat "later assigned to Inspire" as unverified and probably incorrect. This matters practically: the '732 reference being Medtronic's is what supplies the Medtronic 3990 half-cuff electrode link to Goding 1998 (also Medtronic-funded), which is a stronger motivation-to-combine story, not a weaker one.
2. Filing-date ambiguity (already flagged in the prior sections): Google Patents shows 1999‑07‑27 (PCT international filing); Justia shows 2001‑03‑26 (national-stage entry). This does not affect the §103 analysis, because the operative date for prior-art qualification is the 1998‑07‑27 provisional priority date, and every reference relied on above predates it.
3. Specification drafting defect noticed. The '725 specification states "as illustrated in FIG. 8, the interface circuit IC comprises a transformer TR1…" — but the patent has only FIGS. 1–6. This appears in the issued text at the passage corresponding to FIG. 6 and is a typographical/numeral error. It does not change the analysis but is worth noting because a claim-6/12 narrowing argument would rely on the FIG. 6 disclosure.
4. Goding 1998 as §102(a) vs. §102(b). Published Feb. 1998 — less than one year before the 1998‑07‑27 priority date. It therefore qualifies only under pre‑AIA §102(a) (known/published before the invention), not §102(b). If the inventors could swear behind it (Rule 131) or establish an earlier conception/reduction to practice, Goding would drop out. However, the '725 specification itself describes the same canine research program ("Two healthy Beagles … chronically implanted with spiral nerve cuff electrodes 10 on the main trunk of the hypoglossal nerve HG") and reports 53 sleep sessions over 17 months — meaning the '725 work and the Goding/Schwartz/Medtronic work were running concurrently in the same dogs' wake. That coincidence makes a successful swear-behind risky, and it also makes the Testerman/Medtronic art the more reliable §102(b) anchor for the same disclosures.
5. Search limitations. I have no authenticated PACER/CAFC, Lex Machina, or full-text PTAB docket access in this session. Nothing here should be read as a validity adjudication, and the "Expired – Lifetime" legal status is the sources' representation, not a legal conclusion. No party has an interest in testing this claim set, which is precisely why it has never been litigated.
6. No fabrication of references. Every reference relied on above was retrieved and quoted from the sources cited inline. Where a page range or author list was truncated in the search result (e.g., Smith et al., Sleep 1996;19(10 Suppl)), I have signaled that rather than supplying a precise page range.